Loading...
1/*
2 * Copyright (C) 2008 Oracle. All rights reserved.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
17 */
18
19#include <linux/sched.h>
20#include <linux/slab.h>
21#include <linux/blkdev.h>
22#include <linux/list_sort.h>
23#include "tree-log.h"
24#include "disk-io.h"
25#include "locking.h"
26#include "print-tree.h"
27#include "backref.h"
28#include "hash.h"
29#include "compression.h"
30
31/* magic values for the inode_only field in btrfs_log_inode:
32 *
33 * LOG_INODE_ALL means to log everything
34 * LOG_INODE_EXISTS means to log just enough to recreate the inode
35 * during log replay
36 */
37#define LOG_INODE_ALL 0
38#define LOG_INODE_EXISTS 1
39
40/*
41 * directory trouble cases
42 *
43 * 1) on rename or unlink, if the inode being unlinked isn't in the fsync
44 * log, we must force a full commit before doing an fsync of the directory
45 * where the unlink was done.
46 * ---> record transid of last unlink/rename per directory
47 *
48 * mkdir foo/some_dir
49 * normal commit
50 * rename foo/some_dir foo2/some_dir
51 * mkdir foo/some_dir
52 * fsync foo/some_dir/some_file
53 *
54 * The fsync above will unlink the original some_dir without recording
55 * it in its new location (foo2). After a crash, some_dir will be gone
56 * unless the fsync of some_file forces a full commit
57 *
58 * 2) we must log any new names for any file or dir that is in the fsync
59 * log. ---> check inode while renaming/linking.
60 *
61 * 2a) we must log any new names for any file or dir during rename
62 * when the directory they are being removed from was logged.
63 * ---> check inode and old parent dir during rename
64 *
65 * 2a is actually the more important variant. With the extra logging
66 * a crash might unlink the old name without recreating the new one
67 *
68 * 3) after a crash, we must go through any directories with a link count
69 * of zero and redo the rm -rf
70 *
71 * mkdir f1/foo
72 * normal commit
73 * rm -rf f1/foo
74 * fsync(f1)
75 *
76 * The directory f1 was fully removed from the FS, but fsync was never
77 * called on f1, only its parent dir. After a crash the rm -rf must
78 * be replayed. This must be able to recurse down the entire
79 * directory tree. The inode link count fixup code takes care of the
80 * ugly details.
81 */
82
83/*
84 * stages for the tree walking. The first
85 * stage (0) is to only pin down the blocks we find
86 * the second stage (1) is to make sure that all the inodes
87 * we find in the log are created in the subvolume.
88 *
89 * The last stage is to deal with directories and links and extents
90 * and all the other fun semantics
91 */
92#define LOG_WALK_PIN_ONLY 0
93#define LOG_WALK_REPLAY_INODES 1
94#define LOG_WALK_REPLAY_DIR_INDEX 2
95#define LOG_WALK_REPLAY_ALL 3
96
97static int btrfs_log_inode(struct btrfs_trans_handle *trans,
98 struct btrfs_root *root, struct inode *inode,
99 int inode_only,
100 const loff_t start,
101 const loff_t end,
102 struct btrfs_log_ctx *ctx);
103static int link_to_fixup_dir(struct btrfs_trans_handle *trans,
104 struct btrfs_root *root,
105 struct btrfs_path *path, u64 objectid);
106static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
107 struct btrfs_root *root,
108 struct btrfs_root *log,
109 struct btrfs_path *path,
110 u64 dirid, int del_all);
111
112/*
113 * tree logging is a special write ahead log used to make sure that
114 * fsyncs and O_SYNCs can happen without doing full tree commits.
115 *
116 * Full tree commits are expensive because they require commonly
117 * modified blocks to be recowed, creating many dirty pages in the
118 * extent tree an 4x-6x higher write load than ext3.
119 *
120 * Instead of doing a tree commit on every fsync, we use the
121 * key ranges and transaction ids to find items for a given file or directory
122 * that have changed in this transaction. Those items are copied into
123 * a special tree (one per subvolume root), that tree is written to disk
124 * and then the fsync is considered complete.
125 *
126 * After a crash, items are copied out of the log-tree back into the
127 * subvolume tree. Any file data extents found are recorded in the extent
128 * allocation tree, and the log-tree freed.
129 *
130 * The log tree is read three times, once to pin down all the extents it is
131 * using in ram and once, once to create all the inodes logged in the tree
132 * and once to do all the other items.
133 */
134
135/*
136 * start a sub transaction and setup the log tree
137 * this increments the log tree writer count to make the people
138 * syncing the tree wait for us to finish
139 */
140static int start_log_trans(struct btrfs_trans_handle *trans,
141 struct btrfs_root *root,
142 struct btrfs_log_ctx *ctx)
143{
144 int ret = 0;
145
146 mutex_lock(&root->log_mutex);
147
148 if (root->log_root) {
149 if (btrfs_need_log_full_commit(root->fs_info, trans)) {
150 ret = -EAGAIN;
151 goto out;
152 }
153
154 if (!root->log_start_pid) {
155 clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
156 root->log_start_pid = current->pid;
157 } else if (root->log_start_pid != current->pid) {
158 set_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
159 }
160 } else {
161 mutex_lock(&root->fs_info->tree_log_mutex);
162 if (!root->fs_info->log_root_tree)
163 ret = btrfs_init_log_root_tree(trans, root->fs_info);
164 mutex_unlock(&root->fs_info->tree_log_mutex);
165 if (ret)
166 goto out;
167
168 ret = btrfs_add_log_tree(trans, root);
169 if (ret)
170 goto out;
171
172 clear_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state);
173 root->log_start_pid = current->pid;
174 }
175
176 atomic_inc(&root->log_batch);
177 atomic_inc(&root->log_writers);
178 if (ctx) {
179 int index = root->log_transid % 2;
180 list_add_tail(&ctx->list, &root->log_ctxs[index]);
181 ctx->log_transid = root->log_transid;
182 }
183
184out:
185 mutex_unlock(&root->log_mutex);
186 return ret;
187}
188
189/*
190 * returns 0 if there was a log transaction running and we were able
191 * to join, or returns -ENOENT if there were not transactions
192 * in progress
193 */
194static int join_running_log_trans(struct btrfs_root *root)
195{
196 int ret = -ENOENT;
197
198 smp_mb();
199 if (!root->log_root)
200 return -ENOENT;
201
202 mutex_lock(&root->log_mutex);
203 if (root->log_root) {
204 ret = 0;
205 atomic_inc(&root->log_writers);
206 }
207 mutex_unlock(&root->log_mutex);
208 return ret;
209}
210
211/*
212 * This either makes the current running log transaction wait
213 * until you call btrfs_end_log_trans() or it makes any future
214 * log transactions wait until you call btrfs_end_log_trans()
215 */
216int btrfs_pin_log_trans(struct btrfs_root *root)
217{
218 int ret = -ENOENT;
219
220 mutex_lock(&root->log_mutex);
221 atomic_inc(&root->log_writers);
222 mutex_unlock(&root->log_mutex);
223 return ret;
224}
225
226/*
227 * indicate we're done making changes to the log tree
228 * and wake up anyone waiting to do a sync
229 */
230void btrfs_end_log_trans(struct btrfs_root *root)
231{
232 if (atomic_dec_and_test(&root->log_writers)) {
233 /*
234 * Implicit memory barrier after atomic_dec_and_test
235 */
236 if (waitqueue_active(&root->log_writer_wait))
237 wake_up(&root->log_writer_wait);
238 }
239}
240
241
242/*
243 * the walk control struct is used to pass state down the chain when
244 * processing the log tree. The stage field tells us which part
245 * of the log tree processing we are currently doing. The others
246 * are state fields used for that specific part
247 */
248struct walk_control {
249 /* should we free the extent on disk when done? This is used
250 * at transaction commit time while freeing a log tree
251 */
252 int free;
253
254 /* should we write out the extent buffer? This is used
255 * while flushing the log tree to disk during a sync
256 */
257 int write;
258
259 /* should we wait for the extent buffer io to finish? Also used
260 * while flushing the log tree to disk for a sync
261 */
262 int wait;
263
264 /* pin only walk, we record which extents on disk belong to the
265 * log trees
266 */
267 int pin;
268
269 /* what stage of the replay code we're currently in */
270 int stage;
271
272 /* the root we are currently replaying */
273 struct btrfs_root *replay_dest;
274
275 /* the trans handle for the current replay */
276 struct btrfs_trans_handle *trans;
277
278 /* the function that gets used to process blocks we find in the
279 * tree. Note the extent_buffer might not be up to date when it is
280 * passed in, and it must be checked or read if you need the data
281 * inside it
282 */
283 int (*process_func)(struct btrfs_root *log, struct extent_buffer *eb,
284 struct walk_control *wc, u64 gen);
285};
286
287/*
288 * process_func used to pin down extents, write them or wait on them
289 */
290static int process_one_buffer(struct btrfs_root *log,
291 struct extent_buffer *eb,
292 struct walk_control *wc, u64 gen)
293{
294 int ret = 0;
295
296 /*
297 * If this fs is mixed then we need to be able to process the leaves to
298 * pin down any logged extents, so we have to read the block.
299 */
300 if (btrfs_fs_incompat(log->fs_info, MIXED_GROUPS)) {
301 ret = btrfs_read_buffer(eb, gen);
302 if (ret)
303 return ret;
304 }
305
306 if (wc->pin)
307 ret = btrfs_pin_extent_for_log_replay(log->fs_info->extent_root,
308 eb->start, eb->len);
309
310 if (!ret && btrfs_buffer_uptodate(eb, gen, 0)) {
311 if (wc->pin && btrfs_header_level(eb) == 0)
312 ret = btrfs_exclude_logged_extents(log, eb);
313 if (wc->write)
314 btrfs_write_tree_block(eb);
315 if (wc->wait)
316 btrfs_wait_tree_block_writeback(eb);
317 }
318 return ret;
319}
320
321/*
322 * Item overwrite used by replay and tree logging. eb, slot and key all refer
323 * to the src data we are copying out.
324 *
325 * root is the tree we are copying into, and path is a scratch
326 * path for use in this function (it should be released on entry and
327 * will be released on exit).
328 *
329 * If the key is already in the destination tree the existing item is
330 * overwritten. If the existing item isn't big enough, it is extended.
331 * If it is too large, it is truncated.
332 *
333 * If the key isn't in the destination yet, a new item is inserted.
334 */
335static noinline int overwrite_item(struct btrfs_trans_handle *trans,
336 struct btrfs_root *root,
337 struct btrfs_path *path,
338 struct extent_buffer *eb, int slot,
339 struct btrfs_key *key)
340{
341 int ret;
342 u32 item_size;
343 u64 saved_i_size = 0;
344 int save_old_i_size = 0;
345 unsigned long src_ptr;
346 unsigned long dst_ptr;
347 int overwrite_root = 0;
348 bool inode_item = key->type == BTRFS_INODE_ITEM_KEY;
349
350 if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
351 overwrite_root = 1;
352
353 item_size = btrfs_item_size_nr(eb, slot);
354 src_ptr = btrfs_item_ptr_offset(eb, slot);
355
356 /* look for the key in the destination tree */
357 ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
358 if (ret < 0)
359 return ret;
360
361 if (ret == 0) {
362 char *src_copy;
363 char *dst_copy;
364 u32 dst_size = btrfs_item_size_nr(path->nodes[0],
365 path->slots[0]);
366 if (dst_size != item_size)
367 goto insert;
368
369 if (item_size == 0) {
370 btrfs_release_path(path);
371 return 0;
372 }
373 dst_copy = kmalloc(item_size, GFP_NOFS);
374 src_copy = kmalloc(item_size, GFP_NOFS);
375 if (!dst_copy || !src_copy) {
376 btrfs_release_path(path);
377 kfree(dst_copy);
378 kfree(src_copy);
379 return -ENOMEM;
380 }
381
382 read_extent_buffer(eb, src_copy, src_ptr, item_size);
383
384 dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
385 read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
386 item_size);
387 ret = memcmp(dst_copy, src_copy, item_size);
388
389 kfree(dst_copy);
390 kfree(src_copy);
391 /*
392 * they have the same contents, just return, this saves
393 * us from cowing blocks in the destination tree and doing
394 * extra writes that may not have been done by a previous
395 * sync
396 */
397 if (ret == 0) {
398 btrfs_release_path(path);
399 return 0;
400 }
401
402 /*
403 * We need to load the old nbytes into the inode so when we
404 * replay the extents we've logged we get the right nbytes.
405 */
406 if (inode_item) {
407 struct btrfs_inode_item *item;
408 u64 nbytes;
409 u32 mode;
410
411 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
412 struct btrfs_inode_item);
413 nbytes = btrfs_inode_nbytes(path->nodes[0], item);
414 item = btrfs_item_ptr(eb, slot,
415 struct btrfs_inode_item);
416 btrfs_set_inode_nbytes(eb, item, nbytes);
417
418 /*
419 * If this is a directory we need to reset the i_size to
420 * 0 so that we can set it up properly when replaying
421 * the rest of the items in this log.
422 */
423 mode = btrfs_inode_mode(eb, item);
424 if (S_ISDIR(mode))
425 btrfs_set_inode_size(eb, item, 0);
426 }
427 } else if (inode_item) {
428 struct btrfs_inode_item *item;
429 u32 mode;
430
431 /*
432 * New inode, set nbytes to 0 so that the nbytes comes out
433 * properly when we replay the extents.
434 */
435 item = btrfs_item_ptr(eb, slot, struct btrfs_inode_item);
436 btrfs_set_inode_nbytes(eb, item, 0);
437
438 /*
439 * If this is a directory we need to reset the i_size to 0 so
440 * that we can set it up properly when replaying the rest of
441 * the items in this log.
442 */
443 mode = btrfs_inode_mode(eb, item);
444 if (S_ISDIR(mode))
445 btrfs_set_inode_size(eb, item, 0);
446 }
447insert:
448 btrfs_release_path(path);
449 /* try to insert the key into the destination tree */
450 path->skip_release_on_error = 1;
451 ret = btrfs_insert_empty_item(trans, root, path,
452 key, item_size);
453 path->skip_release_on_error = 0;
454
455 /* make sure any existing item is the correct size */
456 if (ret == -EEXIST || ret == -EOVERFLOW) {
457 u32 found_size;
458 found_size = btrfs_item_size_nr(path->nodes[0],
459 path->slots[0]);
460 if (found_size > item_size)
461 btrfs_truncate_item(root, path, item_size, 1);
462 else if (found_size < item_size)
463 btrfs_extend_item(root, path,
464 item_size - found_size);
465 } else if (ret) {
466 return ret;
467 }
468 dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
469 path->slots[0]);
470
471 /* don't overwrite an existing inode if the generation number
472 * was logged as zero. This is done when the tree logging code
473 * is just logging an inode to make sure it exists after recovery.
474 *
475 * Also, don't overwrite i_size on directories during replay.
476 * log replay inserts and removes directory items based on the
477 * state of the tree found in the subvolume, and i_size is modified
478 * as it goes
479 */
480 if (key->type == BTRFS_INODE_ITEM_KEY && ret == -EEXIST) {
481 struct btrfs_inode_item *src_item;
482 struct btrfs_inode_item *dst_item;
483
484 src_item = (struct btrfs_inode_item *)src_ptr;
485 dst_item = (struct btrfs_inode_item *)dst_ptr;
486
487 if (btrfs_inode_generation(eb, src_item) == 0) {
488 struct extent_buffer *dst_eb = path->nodes[0];
489 const u64 ino_size = btrfs_inode_size(eb, src_item);
490
491 /*
492 * For regular files an ino_size == 0 is used only when
493 * logging that an inode exists, as part of a directory
494 * fsync, and the inode wasn't fsynced before. In this
495 * case don't set the size of the inode in the fs/subvol
496 * tree, otherwise we would be throwing valid data away.
497 */
498 if (S_ISREG(btrfs_inode_mode(eb, src_item)) &&
499 S_ISREG(btrfs_inode_mode(dst_eb, dst_item)) &&
500 ino_size != 0) {
501 struct btrfs_map_token token;
502
503 btrfs_init_map_token(&token);
504 btrfs_set_token_inode_size(dst_eb, dst_item,
505 ino_size, &token);
506 }
507 goto no_copy;
508 }
509
510 if (overwrite_root &&
511 S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
512 S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
513 save_old_i_size = 1;
514 saved_i_size = btrfs_inode_size(path->nodes[0],
515 dst_item);
516 }
517 }
518
519 copy_extent_buffer(path->nodes[0], eb, dst_ptr,
520 src_ptr, item_size);
521
522 if (save_old_i_size) {
523 struct btrfs_inode_item *dst_item;
524 dst_item = (struct btrfs_inode_item *)dst_ptr;
525 btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
526 }
527
528 /* make sure the generation is filled in */
529 if (key->type == BTRFS_INODE_ITEM_KEY) {
530 struct btrfs_inode_item *dst_item;
531 dst_item = (struct btrfs_inode_item *)dst_ptr;
532 if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
533 btrfs_set_inode_generation(path->nodes[0], dst_item,
534 trans->transid);
535 }
536 }
537no_copy:
538 btrfs_mark_buffer_dirty(path->nodes[0]);
539 btrfs_release_path(path);
540 return 0;
541}
542
543/*
544 * simple helper to read an inode off the disk from a given root
545 * This can only be called for subvolume roots and not for the log
546 */
547static noinline struct inode *read_one_inode(struct btrfs_root *root,
548 u64 objectid)
549{
550 struct btrfs_key key;
551 struct inode *inode;
552
553 key.objectid = objectid;
554 key.type = BTRFS_INODE_ITEM_KEY;
555 key.offset = 0;
556 inode = btrfs_iget(root->fs_info->sb, &key, root, NULL);
557 if (IS_ERR(inode)) {
558 inode = NULL;
559 } else if (is_bad_inode(inode)) {
560 iput(inode);
561 inode = NULL;
562 }
563 return inode;
564}
565
566/* replays a single extent in 'eb' at 'slot' with 'key' into the
567 * subvolume 'root'. path is released on entry and should be released
568 * on exit.
569 *
570 * extents in the log tree have not been allocated out of the extent
571 * tree yet. So, this completes the allocation, taking a reference
572 * as required if the extent already exists or creating a new extent
573 * if it isn't in the extent allocation tree yet.
574 *
575 * The extent is inserted into the file, dropping any existing extents
576 * from the file that overlap the new one.
577 */
578static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
579 struct btrfs_root *root,
580 struct btrfs_path *path,
581 struct extent_buffer *eb, int slot,
582 struct btrfs_key *key)
583{
584 int found_type;
585 u64 extent_end;
586 u64 start = key->offset;
587 u64 nbytes = 0;
588 struct btrfs_file_extent_item *item;
589 struct inode *inode = NULL;
590 unsigned long size;
591 int ret = 0;
592
593 item = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
594 found_type = btrfs_file_extent_type(eb, item);
595
596 if (found_type == BTRFS_FILE_EXTENT_REG ||
597 found_type == BTRFS_FILE_EXTENT_PREALLOC) {
598 nbytes = btrfs_file_extent_num_bytes(eb, item);
599 extent_end = start + nbytes;
600
601 /*
602 * We don't add to the inodes nbytes if we are prealloc or a
603 * hole.
604 */
605 if (btrfs_file_extent_disk_bytenr(eb, item) == 0)
606 nbytes = 0;
607 } else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
608 size = btrfs_file_extent_inline_len(eb, slot, item);
609 nbytes = btrfs_file_extent_ram_bytes(eb, item);
610 extent_end = ALIGN(start + size, root->sectorsize);
611 } else {
612 ret = 0;
613 goto out;
614 }
615
616 inode = read_one_inode(root, key->objectid);
617 if (!inode) {
618 ret = -EIO;
619 goto out;
620 }
621
622 /*
623 * first check to see if we already have this extent in the
624 * file. This must be done before the btrfs_drop_extents run
625 * so we don't try to drop this extent.
626 */
627 ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
628 start, 0);
629
630 if (ret == 0 &&
631 (found_type == BTRFS_FILE_EXTENT_REG ||
632 found_type == BTRFS_FILE_EXTENT_PREALLOC)) {
633 struct btrfs_file_extent_item cmp1;
634 struct btrfs_file_extent_item cmp2;
635 struct btrfs_file_extent_item *existing;
636 struct extent_buffer *leaf;
637
638 leaf = path->nodes[0];
639 existing = btrfs_item_ptr(leaf, path->slots[0],
640 struct btrfs_file_extent_item);
641
642 read_extent_buffer(eb, &cmp1, (unsigned long)item,
643 sizeof(cmp1));
644 read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
645 sizeof(cmp2));
646
647 /*
648 * we already have a pointer to this exact extent,
649 * we don't have to do anything
650 */
651 if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
652 btrfs_release_path(path);
653 goto out;
654 }
655 }
656 btrfs_release_path(path);
657
658 /* drop any overlapping extents */
659 ret = btrfs_drop_extents(trans, root, inode, start, extent_end, 1);
660 if (ret)
661 goto out;
662
663 if (found_type == BTRFS_FILE_EXTENT_REG ||
664 found_type == BTRFS_FILE_EXTENT_PREALLOC) {
665 u64 offset;
666 unsigned long dest_offset;
667 struct btrfs_key ins;
668
669 ret = btrfs_insert_empty_item(trans, root, path, key,
670 sizeof(*item));
671 if (ret)
672 goto out;
673 dest_offset = btrfs_item_ptr_offset(path->nodes[0],
674 path->slots[0]);
675 copy_extent_buffer(path->nodes[0], eb, dest_offset,
676 (unsigned long)item, sizeof(*item));
677
678 ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
679 ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
680 ins.type = BTRFS_EXTENT_ITEM_KEY;
681 offset = key->offset - btrfs_file_extent_offset(eb, item);
682
683 if (ins.objectid > 0) {
684 u64 csum_start;
685 u64 csum_end;
686 LIST_HEAD(ordered_sums);
687 /*
688 * is this extent already allocated in the extent
689 * allocation tree? If so, just add a reference
690 */
691 ret = btrfs_lookup_data_extent(root, ins.objectid,
692 ins.offset);
693 if (ret == 0) {
694 ret = btrfs_inc_extent_ref(trans, root,
695 ins.objectid, ins.offset,
696 0, root->root_key.objectid,
697 key->objectid, offset);
698 if (ret)
699 goto out;
700 } else {
701 /*
702 * insert the extent pointer in the extent
703 * allocation tree
704 */
705 ret = btrfs_alloc_logged_file_extent(trans,
706 root, root->root_key.objectid,
707 key->objectid, offset, &ins);
708 if (ret)
709 goto out;
710 }
711 btrfs_release_path(path);
712
713 if (btrfs_file_extent_compression(eb, item)) {
714 csum_start = ins.objectid;
715 csum_end = csum_start + ins.offset;
716 } else {
717 csum_start = ins.objectid +
718 btrfs_file_extent_offset(eb, item);
719 csum_end = csum_start +
720 btrfs_file_extent_num_bytes(eb, item);
721 }
722
723 ret = btrfs_lookup_csums_range(root->log_root,
724 csum_start, csum_end - 1,
725 &ordered_sums, 0);
726 if (ret)
727 goto out;
728 /*
729 * Now delete all existing cums in the csum root that
730 * cover our range. We do this because we can have an
731 * extent that is completely referenced by one file
732 * extent item and partially referenced by another
733 * file extent item (like after using the clone or
734 * extent_same ioctls). In this case if we end up doing
735 * the replay of the one that partially references the
736 * extent first, and we do not do the csum deletion
737 * below, we can get 2 csum items in the csum tree that
738 * overlap each other. For example, imagine our log has
739 * the two following file extent items:
740 *
741 * key (257 EXTENT_DATA 409600)
742 * extent data disk byte 12845056 nr 102400
743 * extent data offset 20480 nr 20480 ram 102400
744 *
745 * key (257 EXTENT_DATA 819200)
746 * extent data disk byte 12845056 nr 102400
747 * extent data offset 0 nr 102400 ram 102400
748 *
749 * Where the second one fully references the 100K extent
750 * that starts at disk byte 12845056, and the log tree
751 * has a single csum item that covers the entire range
752 * of the extent:
753 *
754 * key (EXTENT_CSUM EXTENT_CSUM 12845056) itemsize 100
755 *
756 * After the first file extent item is replayed, the
757 * csum tree gets the following csum item:
758 *
759 * key (EXTENT_CSUM EXTENT_CSUM 12865536) itemsize 20
760 *
761 * Which covers the 20K sub-range starting at offset 20K
762 * of our extent. Now when we replay the second file
763 * extent item, if we do not delete existing csum items
764 * that cover any of its blocks, we end up getting two
765 * csum items in our csum tree that overlap each other:
766 *
767 * key (EXTENT_CSUM EXTENT_CSUM 12845056) itemsize 100
768 * key (EXTENT_CSUM EXTENT_CSUM 12865536) itemsize 20
769 *
770 * Which is a problem, because after this anyone trying
771 * to lookup up for the checksum of any block of our
772 * extent starting at an offset of 40K or higher, will
773 * end up looking at the second csum item only, which
774 * does not contain the checksum for any block starting
775 * at offset 40K or higher of our extent.
776 */
777 while (!list_empty(&ordered_sums)) {
778 struct btrfs_ordered_sum *sums;
779 sums = list_entry(ordered_sums.next,
780 struct btrfs_ordered_sum,
781 list);
782 if (!ret)
783 ret = btrfs_del_csums(trans,
784 root->fs_info->csum_root,
785 sums->bytenr,
786 sums->len);
787 if (!ret)
788 ret = btrfs_csum_file_blocks(trans,
789 root->fs_info->csum_root,
790 sums);
791 list_del(&sums->list);
792 kfree(sums);
793 }
794 if (ret)
795 goto out;
796 } else {
797 btrfs_release_path(path);
798 }
799 } else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
800 /* inline extents are easy, we just overwrite them */
801 ret = overwrite_item(trans, root, path, eb, slot, key);
802 if (ret)
803 goto out;
804 }
805
806 inode_add_bytes(inode, nbytes);
807 ret = btrfs_update_inode(trans, root, inode);
808out:
809 if (inode)
810 iput(inode);
811 return ret;
812}
813
814/*
815 * when cleaning up conflicts between the directory names in the
816 * subvolume, directory names in the log and directory names in the
817 * inode back references, we may have to unlink inodes from directories.
818 *
819 * This is a helper function to do the unlink of a specific directory
820 * item
821 */
822static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
823 struct btrfs_root *root,
824 struct btrfs_path *path,
825 struct inode *dir,
826 struct btrfs_dir_item *di)
827{
828 struct inode *inode;
829 char *name;
830 int name_len;
831 struct extent_buffer *leaf;
832 struct btrfs_key location;
833 int ret;
834
835 leaf = path->nodes[0];
836
837 btrfs_dir_item_key_to_cpu(leaf, di, &location);
838 name_len = btrfs_dir_name_len(leaf, di);
839 name = kmalloc(name_len, GFP_NOFS);
840 if (!name)
841 return -ENOMEM;
842
843 read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
844 btrfs_release_path(path);
845
846 inode = read_one_inode(root, location.objectid);
847 if (!inode) {
848 ret = -EIO;
849 goto out;
850 }
851
852 ret = link_to_fixup_dir(trans, root, path, location.objectid);
853 if (ret)
854 goto out;
855
856 ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
857 if (ret)
858 goto out;
859 else
860 ret = btrfs_run_delayed_items(trans, root);
861out:
862 kfree(name);
863 iput(inode);
864 return ret;
865}
866
867/*
868 * helper function to see if a given name and sequence number found
869 * in an inode back reference are already in a directory and correctly
870 * point to this inode
871 */
872static noinline int inode_in_dir(struct btrfs_root *root,
873 struct btrfs_path *path,
874 u64 dirid, u64 objectid, u64 index,
875 const char *name, int name_len)
876{
877 struct btrfs_dir_item *di;
878 struct btrfs_key location;
879 int match = 0;
880
881 di = btrfs_lookup_dir_index_item(NULL, root, path, dirid,
882 index, name, name_len, 0);
883 if (di && !IS_ERR(di)) {
884 btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
885 if (location.objectid != objectid)
886 goto out;
887 } else
888 goto out;
889 btrfs_release_path(path);
890
891 di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
892 if (di && !IS_ERR(di)) {
893 btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
894 if (location.objectid != objectid)
895 goto out;
896 } else
897 goto out;
898 match = 1;
899out:
900 btrfs_release_path(path);
901 return match;
902}
903
904/*
905 * helper function to check a log tree for a named back reference in
906 * an inode. This is used to decide if a back reference that is
907 * found in the subvolume conflicts with what we find in the log.
908 *
909 * inode backreferences may have multiple refs in a single item,
910 * during replay we process one reference at a time, and we don't
911 * want to delete valid links to a file from the subvolume if that
912 * link is also in the log.
913 */
914static noinline int backref_in_log(struct btrfs_root *log,
915 struct btrfs_key *key,
916 u64 ref_objectid,
917 const char *name, int namelen)
918{
919 struct btrfs_path *path;
920 struct btrfs_inode_ref *ref;
921 unsigned long ptr;
922 unsigned long ptr_end;
923 unsigned long name_ptr;
924 int found_name_len;
925 int item_size;
926 int ret;
927 int match = 0;
928
929 path = btrfs_alloc_path();
930 if (!path)
931 return -ENOMEM;
932
933 ret = btrfs_search_slot(NULL, log, key, path, 0, 0);
934 if (ret != 0)
935 goto out;
936
937 ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
938
939 if (key->type == BTRFS_INODE_EXTREF_KEY) {
940 if (btrfs_find_name_in_ext_backref(path, ref_objectid,
941 name, namelen, NULL))
942 match = 1;
943
944 goto out;
945 }
946
947 item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
948 ptr_end = ptr + item_size;
949 while (ptr < ptr_end) {
950 ref = (struct btrfs_inode_ref *)ptr;
951 found_name_len = btrfs_inode_ref_name_len(path->nodes[0], ref);
952 if (found_name_len == namelen) {
953 name_ptr = (unsigned long)(ref + 1);
954 ret = memcmp_extent_buffer(path->nodes[0], name,
955 name_ptr, namelen);
956 if (ret == 0) {
957 match = 1;
958 goto out;
959 }
960 }
961 ptr = (unsigned long)(ref + 1) + found_name_len;
962 }
963out:
964 btrfs_free_path(path);
965 return match;
966}
967
968static inline int __add_inode_ref(struct btrfs_trans_handle *trans,
969 struct btrfs_root *root,
970 struct btrfs_path *path,
971 struct btrfs_root *log_root,
972 struct inode *dir, struct inode *inode,
973 struct extent_buffer *eb,
974 u64 inode_objectid, u64 parent_objectid,
975 u64 ref_index, char *name, int namelen,
976 int *search_done)
977{
978 int ret;
979 char *victim_name;
980 int victim_name_len;
981 struct extent_buffer *leaf;
982 struct btrfs_dir_item *di;
983 struct btrfs_key search_key;
984 struct btrfs_inode_extref *extref;
985
986again:
987 /* Search old style refs */
988 search_key.objectid = inode_objectid;
989 search_key.type = BTRFS_INODE_REF_KEY;
990 search_key.offset = parent_objectid;
991 ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
992 if (ret == 0) {
993 struct btrfs_inode_ref *victim_ref;
994 unsigned long ptr;
995 unsigned long ptr_end;
996
997 leaf = path->nodes[0];
998
999 /* are we trying to overwrite a back ref for the root directory
1000 * if so, just jump out, we're done
1001 */
1002 if (search_key.objectid == search_key.offset)
1003 return 1;
1004
1005 /* check all the names in this back reference to see
1006 * if they are in the log. if so, we allow them to stay
1007 * otherwise they must be unlinked as a conflict
1008 */
1009 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1010 ptr_end = ptr + btrfs_item_size_nr(leaf, path->slots[0]);
1011 while (ptr < ptr_end) {
1012 victim_ref = (struct btrfs_inode_ref *)ptr;
1013 victim_name_len = btrfs_inode_ref_name_len(leaf,
1014 victim_ref);
1015 victim_name = kmalloc(victim_name_len, GFP_NOFS);
1016 if (!victim_name)
1017 return -ENOMEM;
1018
1019 read_extent_buffer(leaf, victim_name,
1020 (unsigned long)(victim_ref + 1),
1021 victim_name_len);
1022
1023 if (!backref_in_log(log_root, &search_key,
1024 parent_objectid,
1025 victim_name,
1026 victim_name_len)) {
1027 inc_nlink(inode);
1028 btrfs_release_path(path);
1029
1030 ret = btrfs_unlink_inode(trans, root, dir,
1031 inode, victim_name,
1032 victim_name_len);
1033 kfree(victim_name);
1034 if (ret)
1035 return ret;
1036 ret = btrfs_run_delayed_items(trans, root);
1037 if (ret)
1038 return ret;
1039 *search_done = 1;
1040 goto again;
1041 }
1042 kfree(victim_name);
1043
1044 ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
1045 }
1046
1047 /*
1048 * NOTE: we have searched root tree and checked the
1049 * corresponding ref, it does not need to check again.
1050 */
1051 *search_done = 1;
1052 }
1053 btrfs_release_path(path);
1054
1055 /* Same search but for extended refs */
1056 extref = btrfs_lookup_inode_extref(NULL, root, path, name, namelen,
1057 inode_objectid, parent_objectid, 0,
1058 0);
1059 if (!IS_ERR_OR_NULL(extref)) {
1060 u32 item_size;
1061 u32 cur_offset = 0;
1062 unsigned long base;
1063 struct inode *victim_parent;
1064
1065 leaf = path->nodes[0];
1066
1067 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1068 base = btrfs_item_ptr_offset(leaf, path->slots[0]);
1069
1070 while (cur_offset < item_size) {
1071 extref = (struct btrfs_inode_extref *)(base + cur_offset);
1072
1073 victim_name_len = btrfs_inode_extref_name_len(leaf, extref);
1074
1075 if (btrfs_inode_extref_parent(leaf, extref) != parent_objectid)
1076 goto next;
1077
1078 victim_name = kmalloc(victim_name_len, GFP_NOFS);
1079 if (!victim_name)
1080 return -ENOMEM;
1081 read_extent_buffer(leaf, victim_name, (unsigned long)&extref->name,
1082 victim_name_len);
1083
1084 search_key.objectid = inode_objectid;
1085 search_key.type = BTRFS_INODE_EXTREF_KEY;
1086 search_key.offset = btrfs_extref_hash(parent_objectid,
1087 victim_name,
1088 victim_name_len);
1089 ret = 0;
1090 if (!backref_in_log(log_root, &search_key,
1091 parent_objectid, victim_name,
1092 victim_name_len)) {
1093 ret = -ENOENT;
1094 victim_parent = read_one_inode(root,
1095 parent_objectid);
1096 if (victim_parent) {
1097 inc_nlink(inode);
1098 btrfs_release_path(path);
1099
1100 ret = btrfs_unlink_inode(trans, root,
1101 victim_parent,
1102 inode,
1103 victim_name,
1104 victim_name_len);
1105 if (!ret)
1106 ret = btrfs_run_delayed_items(
1107 trans, root);
1108 }
1109 iput(victim_parent);
1110 kfree(victim_name);
1111 if (ret)
1112 return ret;
1113 *search_done = 1;
1114 goto again;
1115 }
1116 kfree(victim_name);
1117 if (ret)
1118 return ret;
1119next:
1120 cur_offset += victim_name_len + sizeof(*extref);
1121 }
1122 *search_done = 1;
1123 }
1124 btrfs_release_path(path);
1125
1126 /* look for a conflicting sequence number */
1127 di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
1128 ref_index, name, namelen, 0);
1129 if (di && !IS_ERR(di)) {
1130 ret = drop_one_dir_item(trans, root, path, dir, di);
1131 if (ret)
1132 return ret;
1133 }
1134 btrfs_release_path(path);
1135
1136 /* look for a conflicing name */
1137 di = btrfs_lookup_dir_item(trans, root, path, btrfs_ino(dir),
1138 name, namelen, 0);
1139 if (di && !IS_ERR(di)) {
1140 ret = drop_one_dir_item(trans, root, path, dir, di);
1141 if (ret)
1142 return ret;
1143 }
1144 btrfs_release_path(path);
1145
1146 return 0;
1147}
1148
1149static int extref_get_fields(struct extent_buffer *eb, unsigned long ref_ptr,
1150 u32 *namelen, char **name, u64 *index,
1151 u64 *parent_objectid)
1152{
1153 struct btrfs_inode_extref *extref;
1154
1155 extref = (struct btrfs_inode_extref *)ref_ptr;
1156
1157 *namelen = btrfs_inode_extref_name_len(eb, extref);
1158 *name = kmalloc(*namelen, GFP_NOFS);
1159 if (*name == NULL)
1160 return -ENOMEM;
1161
1162 read_extent_buffer(eb, *name, (unsigned long)&extref->name,
1163 *namelen);
1164
1165 *index = btrfs_inode_extref_index(eb, extref);
1166 if (parent_objectid)
1167 *parent_objectid = btrfs_inode_extref_parent(eb, extref);
1168
1169 return 0;
1170}
1171
1172static int ref_get_fields(struct extent_buffer *eb, unsigned long ref_ptr,
1173 u32 *namelen, char **name, u64 *index)
1174{
1175 struct btrfs_inode_ref *ref;
1176
1177 ref = (struct btrfs_inode_ref *)ref_ptr;
1178
1179 *namelen = btrfs_inode_ref_name_len(eb, ref);
1180 *name = kmalloc(*namelen, GFP_NOFS);
1181 if (*name == NULL)
1182 return -ENOMEM;
1183
1184 read_extent_buffer(eb, *name, (unsigned long)(ref + 1), *namelen);
1185
1186 *index = btrfs_inode_ref_index(eb, ref);
1187
1188 return 0;
1189}
1190
1191/*
1192 * replay one inode back reference item found in the log tree.
1193 * eb, slot and key refer to the buffer and key found in the log tree.
1194 * root is the destination we are replaying into, and path is for temp
1195 * use by this function. (it should be released on return).
1196 */
1197static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
1198 struct btrfs_root *root,
1199 struct btrfs_root *log,
1200 struct btrfs_path *path,
1201 struct extent_buffer *eb, int slot,
1202 struct btrfs_key *key)
1203{
1204 struct inode *dir = NULL;
1205 struct inode *inode = NULL;
1206 unsigned long ref_ptr;
1207 unsigned long ref_end;
1208 char *name = NULL;
1209 int namelen;
1210 int ret;
1211 int search_done = 0;
1212 int log_ref_ver = 0;
1213 u64 parent_objectid;
1214 u64 inode_objectid;
1215 u64 ref_index = 0;
1216 int ref_struct_size;
1217
1218 ref_ptr = btrfs_item_ptr_offset(eb, slot);
1219 ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
1220
1221 if (key->type == BTRFS_INODE_EXTREF_KEY) {
1222 struct btrfs_inode_extref *r;
1223
1224 ref_struct_size = sizeof(struct btrfs_inode_extref);
1225 log_ref_ver = 1;
1226 r = (struct btrfs_inode_extref *)ref_ptr;
1227 parent_objectid = btrfs_inode_extref_parent(eb, r);
1228 } else {
1229 ref_struct_size = sizeof(struct btrfs_inode_ref);
1230 parent_objectid = key->offset;
1231 }
1232 inode_objectid = key->objectid;
1233
1234 /*
1235 * it is possible that we didn't log all the parent directories
1236 * for a given inode. If we don't find the dir, just don't
1237 * copy the back ref in. The link count fixup code will take
1238 * care of the rest
1239 */
1240 dir = read_one_inode(root, parent_objectid);
1241 if (!dir) {
1242 ret = -ENOENT;
1243 goto out;
1244 }
1245
1246 inode = read_one_inode(root, inode_objectid);
1247 if (!inode) {
1248 ret = -EIO;
1249 goto out;
1250 }
1251
1252 while (ref_ptr < ref_end) {
1253 if (log_ref_ver) {
1254 ret = extref_get_fields(eb, ref_ptr, &namelen, &name,
1255 &ref_index, &parent_objectid);
1256 /*
1257 * parent object can change from one array
1258 * item to another.
1259 */
1260 if (!dir)
1261 dir = read_one_inode(root, parent_objectid);
1262 if (!dir) {
1263 ret = -ENOENT;
1264 goto out;
1265 }
1266 } else {
1267 ret = ref_get_fields(eb, ref_ptr, &namelen, &name,
1268 &ref_index);
1269 }
1270 if (ret)
1271 goto out;
1272
1273 /* if we already have a perfect match, we're done */
1274 if (!inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
1275 ref_index, name, namelen)) {
1276 /*
1277 * look for a conflicting back reference in the
1278 * metadata. if we find one we have to unlink that name
1279 * of the file before we add our new link. Later on, we
1280 * overwrite any existing back reference, and we don't
1281 * want to create dangling pointers in the directory.
1282 */
1283
1284 if (!search_done) {
1285 ret = __add_inode_ref(trans, root, path, log,
1286 dir, inode, eb,
1287 inode_objectid,
1288 parent_objectid,
1289 ref_index, name, namelen,
1290 &search_done);
1291 if (ret) {
1292 if (ret == 1)
1293 ret = 0;
1294 goto out;
1295 }
1296 }
1297
1298 /* insert our name */
1299 ret = btrfs_add_link(trans, dir, inode, name, namelen,
1300 0, ref_index);
1301 if (ret)
1302 goto out;
1303
1304 btrfs_update_inode(trans, root, inode);
1305 }
1306
1307 ref_ptr = (unsigned long)(ref_ptr + ref_struct_size) + namelen;
1308 kfree(name);
1309 name = NULL;
1310 if (log_ref_ver) {
1311 iput(dir);
1312 dir = NULL;
1313 }
1314 }
1315
1316 /* finally write the back reference in the inode */
1317 ret = overwrite_item(trans, root, path, eb, slot, key);
1318out:
1319 btrfs_release_path(path);
1320 kfree(name);
1321 iput(dir);
1322 iput(inode);
1323 return ret;
1324}
1325
1326static int insert_orphan_item(struct btrfs_trans_handle *trans,
1327 struct btrfs_root *root, u64 ino)
1328{
1329 int ret;
1330
1331 ret = btrfs_insert_orphan_item(trans, root, ino);
1332 if (ret == -EEXIST)
1333 ret = 0;
1334
1335 return ret;
1336}
1337
1338static int count_inode_extrefs(struct btrfs_root *root,
1339 struct inode *inode, struct btrfs_path *path)
1340{
1341 int ret = 0;
1342 int name_len;
1343 unsigned int nlink = 0;
1344 u32 item_size;
1345 u32 cur_offset = 0;
1346 u64 inode_objectid = btrfs_ino(inode);
1347 u64 offset = 0;
1348 unsigned long ptr;
1349 struct btrfs_inode_extref *extref;
1350 struct extent_buffer *leaf;
1351
1352 while (1) {
1353 ret = btrfs_find_one_extref(root, inode_objectid, offset, path,
1354 &extref, &offset);
1355 if (ret)
1356 break;
1357
1358 leaf = path->nodes[0];
1359 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1360 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1361 cur_offset = 0;
1362
1363 while (cur_offset < item_size) {
1364 extref = (struct btrfs_inode_extref *) (ptr + cur_offset);
1365 name_len = btrfs_inode_extref_name_len(leaf, extref);
1366
1367 nlink++;
1368
1369 cur_offset += name_len + sizeof(*extref);
1370 }
1371
1372 offset++;
1373 btrfs_release_path(path);
1374 }
1375 btrfs_release_path(path);
1376
1377 if (ret < 0 && ret != -ENOENT)
1378 return ret;
1379 return nlink;
1380}
1381
1382static int count_inode_refs(struct btrfs_root *root,
1383 struct inode *inode, struct btrfs_path *path)
1384{
1385 int ret;
1386 struct btrfs_key key;
1387 unsigned int nlink = 0;
1388 unsigned long ptr;
1389 unsigned long ptr_end;
1390 int name_len;
1391 u64 ino = btrfs_ino(inode);
1392
1393 key.objectid = ino;
1394 key.type = BTRFS_INODE_REF_KEY;
1395 key.offset = (u64)-1;
1396
1397 while (1) {
1398 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1399 if (ret < 0)
1400 break;
1401 if (ret > 0) {
1402 if (path->slots[0] == 0)
1403 break;
1404 path->slots[0]--;
1405 }
1406process_slot:
1407 btrfs_item_key_to_cpu(path->nodes[0], &key,
1408 path->slots[0]);
1409 if (key.objectid != ino ||
1410 key.type != BTRFS_INODE_REF_KEY)
1411 break;
1412 ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
1413 ptr_end = ptr + btrfs_item_size_nr(path->nodes[0],
1414 path->slots[0]);
1415 while (ptr < ptr_end) {
1416 struct btrfs_inode_ref *ref;
1417
1418 ref = (struct btrfs_inode_ref *)ptr;
1419 name_len = btrfs_inode_ref_name_len(path->nodes[0],
1420 ref);
1421 ptr = (unsigned long)(ref + 1) + name_len;
1422 nlink++;
1423 }
1424
1425 if (key.offset == 0)
1426 break;
1427 if (path->slots[0] > 0) {
1428 path->slots[0]--;
1429 goto process_slot;
1430 }
1431 key.offset--;
1432 btrfs_release_path(path);
1433 }
1434 btrfs_release_path(path);
1435
1436 return nlink;
1437}
1438
1439/*
1440 * There are a few corners where the link count of the file can't
1441 * be properly maintained during replay. So, instead of adding
1442 * lots of complexity to the log code, we just scan the backrefs
1443 * for any file that has been through replay.
1444 *
1445 * The scan will update the link count on the inode to reflect the
1446 * number of back refs found. If it goes down to zero, the iput
1447 * will free the inode.
1448 */
1449static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
1450 struct btrfs_root *root,
1451 struct inode *inode)
1452{
1453 struct btrfs_path *path;
1454 int ret;
1455 u64 nlink = 0;
1456 u64 ino = btrfs_ino(inode);
1457
1458 path = btrfs_alloc_path();
1459 if (!path)
1460 return -ENOMEM;
1461
1462 ret = count_inode_refs(root, inode, path);
1463 if (ret < 0)
1464 goto out;
1465
1466 nlink = ret;
1467
1468 ret = count_inode_extrefs(root, inode, path);
1469 if (ret < 0)
1470 goto out;
1471
1472 nlink += ret;
1473
1474 ret = 0;
1475
1476 if (nlink != inode->i_nlink) {
1477 set_nlink(inode, nlink);
1478 btrfs_update_inode(trans, root, inode);
1479 }
1480 BTRFS_I(inode)->index_cnt = (u64)-1;
1481
1482 if (inode->i_nlink == 0) {
1483 if (S_ISDIR(inode->i_mode)) {
1484 ret = replay_dir_deletes(trans, root, NULL, path,
1485 ino, 1);
1486 if (ret)
1487 goto out;
1488 }
1489 ret = insert_orphan_item(trans, root, ino);
1490 }
1491
1492out:
1493 btrfs_free_path(path);
1494 return ret;
1495}
1496
1497static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
1498 struct btrfs_root *root,
1499 struct btrfs_path *path)
1500{
1501 int ret;
1502 struct btrfs_key key;
1503 struct inode *inode;
1504
1505 key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
1506 key.type = BTRFS_ORPHAN_ITEM_KEY;
1507 key.offset = (u64)-1;
1508 while (1) {
1509 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1510 if (ret < 0)
1511 break;
1512
1513 if (ret == 1) {
1514 if (path->slots[0] == 0)
1515 break;
1516 path->slots[0]--;
1517 }
1518
1519 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1520 if (key.objectid != BTRFS_TREE_LOG_FIXUP_OBJECTID ||
1521 key.type != BTRFS_ORPHAN_ITEM_KEY)
1522 break;
1523
1524 ret = btrfs_del_item(trans, root, path);
1525 if (ret)
1526 goto out;
1527
1528 btrfs_release_path(path);
1529 inode = read_one_inode(root, key.offset);
1530 if (!inode)
1531 return -EIO;
1532
1533 ret = fixup_inode_link_count(trans, root, inode);
1534 iput(inode);
1535 if (ret)
1536 goto out;
1537
1538 /*
1539 * fixup on a directory may create new entries,
1540 * make sure we always look for the highset possible
1541 * offset
1542 */
1543 key.offset = (u64)-1;
1544 }
1545 ret = 0;
1546out:
1547 btrfs_release_path(path);
1548 return ret;
1549}
1550
1551
1552/*
1553 * record a given inode in the fixup dir so we can check its link
1554 * count when replay is done. The link count is incremented here
1555 * so the inode won't go away until we check it
1556 */
1557static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
1558 struct btrfs_root *root,
1559 struct btrfs_path *path,
1560 u64 objectid)
1561{
1562 struct btrfs_key key;
1563 int ret = 0;
1564 struct inode *inode;
1565
1566 inode = read_one_inode(root, objectid);
1567 if (!inode)
1568 return -EIO;
1569
1570 key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
1571 key.type = BTRFS_ORPHAN_ITEM_KEY;
1572 key.offset = objectid;
1573
1574 ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
1575
1576 btrfs_release_path(path);
1577 if (ret == 0) {
1578 if (!inode->i_nlink)
1579 set_nlink(inode, 1);
1580 else
1581 inc_nlink(inode);
1582 ret = btrfs_update_inode(trans, root, inode);
1583 } else if (ret == -EEXIST) {
1584 ret = 0;
1585 } else {
1586 BUG(); /* Logic Error */
1587 }
1588 iput(inode);
1589
1590 return ret;
1591}
1592
1593/*
1594 * when replaying the log for a directory, we only insert names
1595 * for inodes that actually exist. This means an fsync on a directory
1596 * does not implicitly fsync all the new files in it
1597 */
1598static noinline int insert_one_name(struct btrfs_trans_handle *trans,
1599 struct btrfs_root *root,
1600 u64 dirid, u64 index,
1601 char *name, int name_len,
1602 struct btrfs_key *location)
1603{
1604 struct inode *inode;
1605 struct inode *dir;
1606 int ret;
1607
1608 inode = read_one_inode(root, location->objectid);
1609 if (!inode)
1610 return -ENOENT;
1611
1612 dir = read_one_inode(root, dirid);
1613 if (!dir) {
1614 iput(inode);
1615 return -EIO;
1616 }
1617
1618 ret = btrfs_add_link(trans, dir, inode, name, name_len, 1, index);
1619
1620 /* FIXME, put inode into FIXUP list */
1621
1622 iput(inode);
1623 iput(dir);
1624 return ret;
1625}
1626
1627/*
1628 * Return true if an inode reference exists in the log for the given name,
1629 * inode and parent inode.
1630 */
1631static bool name_in_log_ref(struct btrfs_root *log_root,
1632 const char *name, const int name_len,
1633 const u64 dirid, const u64 ino)
1634{
1635 struct btrfs_key search_key;
1636
1637 search_key.objectid = ino;
1638 search_key.type = BTRFS_INODE_REF_KEY;
1639 search_key.offset = dirid;
1640 if (backref_in_log(log_root, &search_key, dirid, name, name_len))
1641 return true;
1642
1643 search_key.type = BTRFS_INODE_EXTREF_KEY;
1644 search_key.offset = btrfs_extref_hash(dirid, name, name_len);
1645 if (backref_in_log(log_root, &search_key, dirid, name, name_len))
1646 return true;
1647
1648 return false;
1649}
1650
1651/*
1652 * take a single entry in a log directory item and replay it into
1653 * the subvolume.
1654 *
1655 * if a conflicting item exists in the subdirectory already,
1656 * the inode it points to is unlinked and put into the link count
1657 * fix up tree.
1658 *
1659 * If a name from the log points to a file or directory that does
1660 * not exist in the FS, it is skipped. fsyncs on directories
1661 * do not force down inodes inside that directory, just changes to the
1662 * names or unlinks in a directory.
1663 *
1664 * Returns < 0 on error, 0 if the name wasn't replayed (dentry points to a
1665 * non-existing inode) and 1 if the name was replayed.
1666 */
1667static noinline int replay_one_name(struct btrfs_trans_handle *trans,
1668 struct btrfs_root *root,
1669 struct btrfs_path *path,
1670 struct extent_buffer *eb,
1671 struct btrfs_dir_item *di,
1672 struct btrfs_key *key)
1673{
1674 char *name;
1675 int name_len;
1676 struct btrfs_dir_item *dst_di;
1677 struct btrfs_key found_key;
1678 struct btrfs_key log_key;
1679 struct inode *dir;
1680 u8 log_type;
1681 int exists;
1682 int ret = 0;
1683 bool update_size = (key->type == BTRFS_DIR_INDEX_KEY);
1684 bool name_added = false;
1685
1686 dir = read_one_inode(root, key->objectid);
1687 if (!dir)
1688 return -EIO;
1689
1690 name_len = btrfs_dir_name_len(eb, di);
1691 name = kmalloc(name_len, GFP_NOFS);
1692 if (!name) {
1693 ret = -ENOMEM;
1694 goto out;
1695 }
1696
1697 log_type = btrfs_dir_type(eb, di);
1698 read_extent_buffer(eb, name, (unsigned long)(di + 1),
1699 name_len);
1700
1701 btrfs_dir_item_key_to_cpu(eb, di, &log_key);
1702 exists = btrfs_lookup_inode(trans, root, path, &log_key, 0);
1703 if (exists == 0)
1704 exists = 1;
1705 else
1706 exists = 0;
1707 btrfs_release_path(path);
1708
1709 if (key->type == BTRFS_DIR_ITEM_KEY) {
1710 dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
1711 name, name_len, 1);
1712 } else if (key->type == BTRFS_DIR_INDEX_KEY) {
1713 dst_di = btrfs_lookup_dir_index_item(trans, root, path,
1714 key->objectid,
1715 key->offset, name,
1716 name_len, 1);
1717 } else {
1718 /* Corruption */
1719 ret = -EINVAL;
1720 goto out;
1721 }
1722 if (IS_ERR_OR_NULL(dst_di)) {
1723 /* we need a sequence number to insert, so we only
1724 * do inserts for the BTRFS_DIR_INDEX_KEY types
1725 */
1726 if (key->type != BTRFS_DIR_INDEX_KEY)
1727 goto out;
1728 goto insert;
1729 }
1730
1731 btrfs_dir_item_key_to_cpu(path->nodes[0], dst_di, &found_key);
1732 /* the existing item matches the logged item */
1733 if (found_key.objectid == log_key.objectid &&
1734 found_key.type == log_key.type &&
1735 found_key.offset == log_key.offset &&
1736 btrfs_dir_type(path->nodes[0], dst_di) == log_type) {
1737 update_size = false;
1738 goto out;
1739 }
1740
1741 /*
1742 * don't drop the conflicting directory entry if the inode
1743 * for the new entry doesn't exist
1744 */
1745 if (!exists)
1746 goto out;
1747
1748 ret = drop_one_dir_item(trans, root, path, dir, dst_di);
1749 if (ret)
1750 goto out;
1751
1752 if (key->type == BTRFS_DIR_INDEX_KEY)
1753 goto insert;
1754out:
1755 btrfs_release_path(path);
1756 if (!ret && update_size) {
1757 btrfs_i_size_write(dir, dir->i_size + name_len * 2);
1758 ret = btrfs_update_inode(trans, root, dir);
1759 }
1760 kfree(name);
1761 iput(dir);
1762 if (!ret && name_added)
1763 ret = 1;
1764 return ret;
1765
1766insert:
1767 if (name_in_log_ref(root->log_root, name, name_len,
1768 key->objectid, log_key.objectid)) {
1769 /* The dentry will be added later. */
1770 ret = 0;
1771 update_size = false;
1772 goto out;
1773 }
1774 btrfs_release_path(path);
1775 ret = insert_one_name(trans, root, key->objectid, key->offset,
1776 name, name_len, &log_key);
1777 if (ret && ret != -ENOENT && ret != -EEXIST)
1778 goto out;
1779 if (!ret)
1780 name_added = true;
1781 update_size = false;
1782 ret = 0;
1783 goto out;
1784}
1785
1786/*
1787 * find all the names in a directory item and reconcile them into
1788 * the subvolume. Only BTRFS_DIR_ITEM_KEY types will have more than
1789 * one name in a directory item, but the same code gets used for
1790 * both directory index types
1791 */
1792static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
1793 struct btrfs_root *root,
1794 struct btrfs_path *path,
1795 struct extent_buffer *eb, int slot,
1796 struct btrfs_key *key)
1797{
1798 int ret = 0;
1799 u32 item_size = btrfs_item_size_nr(eb, slot);
1800 struct btrfs_dir_item *di;
1801 int name_len;
1802 unsigned long ptr;
1803 unsigned long ptr_end;
1804 struct btrfs_path *fixup_path = NULL;
1805
1806 ptr = btrfs_item_ptr_offset(eb, slot);
1807 ptr_end = ptr + item_size;
1808 while (ptr < ptr_end) {
1809 di = (struct btrfs_dir_item *)ptr;
1810 if (verify_dir_item(root, eb, di))
1811 return -EIO;
1812 name_len = btrfs_dir_name_len(eb, di);
1813 ret = replay_one_name(trans, root, path, eb, di, key);
1814 if (ret < 0)
1815 break;
1816 ptr = (unsigned long)(di + 1);
1817 ptr += name_len;
1818
1819 /*
1820 * If this entry refers to a non-directory (directories can not
1821 * have a link count > 1) and it was added in the transaction
1822 * that was not committed, make sure we fixup the link count of
1823 * the inode it the entry points to. Otherwise something like
1824 * the following would result in a directory pointing to an
1825 * inode with a wrong link that does not account for this dir
1826 * entry:
1827 *
1828 * mkdir testdir
1829 * touch testdir/foo
1830 * touch testdir/bar
1831 * sync
1832 *
1833 * ln testdir/bar testdir/bar_link
1834 * ln testdir/foo testdir/foo_link
1835 * xfs_io -c "fsync" testdir/bar
1836 *
1837 * <power failure>
1838 *
1839 * mount fs, log replay happens
1840 *
1841 * File foo would remain with a link count of 1 when it has two
1842 * entries pointing to it in the directory testdir. This would
1843 * make it impossible to ever delete the parent directory has
1844 * it would result in stale dentries that can never be deleted.
1845 */
1846 if (ret == 1 && btrfs_dir_type(eb, di) != BTRFS_FT_DIR) {
1847 struct btrfs_key di_key;
1848
1849 if (!fixup_path) {
1850 fixup_path = btrfs_alloc_path();
1851 if (!fixup_path) {
1852 ret = -ENOMEM;
1853 break;
1854 }
1855 }
1856
1857 btrfs_dir_item_key_to_cpu(eb, di, &di_key);
1858 ret = link_to_fixup_dir(trans, root, fixup_path,
1859 di_key.objectid);
1860 if (ret)
1861 break;
1862 }
1863 ret = 0;
1864 }
1865 btrfs_free_path(fixup_path);
1866 return ret;
1867}
1868
1869/*
1870 * directory replay has two parts. There are the standard directory
1871 * items in the log copied from the subvolume, and range items
1872 * created in the log while the subvolume was logged.
1873 *
1874 * The range items tell us which parts of the key space the log
1875 * is authoritative for. During replay, if a key in the subvolume
1876 * directory is in a logged range item, but not actually in the log
1877 * that means it was deleted from the directory before the fsync
1878 * and should be removed.
1879 */
1880static noinline int find_dir_range(struct btrfs_root *root,
1881 struct btrfs_path *path,
1882 u64 dirid, int key_type,
1883 u64 *start_ret, u64 *end_ret)
1884{
1885 struct btrfs_key key;
1886 u64 found_end;
1887 struct btrfs_dir_log_item *item;
1888 int ret;
1889 int nritems;
1890
1891 if (*start_ret == (u64)-1)
1892 return 1;
1893
1894 key.objectid = dirid;
1895 key.type = key_type;
1896 key.offset = *start_ret;
1897
1898 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1899 if (ret < 0)
1900 goto out;
1901 if (ret > 0) {
1902 if (path->slots[0] == 0)
1903 goto out;
1904 path->slots[0]--;
1905 }
1906 if (ret != 0)
1907 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1908
1909 if (key.type != key_type || key.objectid != dirid) {
1910 ret = 1;
1911 goto next;
1912 }
1913 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
1914 struct btrfs_dir_log_item);
1915 found_end = btrfs_dir_log_end(path->nodes[0], item);
1916
1917 if (*start_ret >= key.offset && *start_ret <= found_end) {
1918 ret = 0;
1919 *start_ret = key.offset;
1920 *end_ret = found_end;
1921 goto out;
1922 }
1923 ret = 1;
1924next:
1925 /* check the next slot in the tree to see if it is a valid item */
1926 nritems = btrfs_header_nritems(path->nodes[0]);
1927 if (path->slots[0] >= nritems) {
1928 ret = btrfs_next_leaf(root, path);
1929 if (ret)
1930 goto out;
1931 } else {
1932 path->slots[0]++;
1933 }
1934
1935 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1936
1937 if (key.type != key_type || key.objectid != dirid) {
1938 ret = 1;
1939 goto out;
1940 }
1941 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
1942 struct btrfs_dir_log_item);
1943 found_end = btrfs_dir_log_end(path->nodes[0], item);
1944 *start_ret = key.offset;
1945 *end_ret = found_end;
1946 ret = 0;
1947out:
1948 btrfs_release_path(path);
1949 return ret;
1950}
1951
1952/*
1953 * this looks for a given directory item in the log. If the directory
1954 * item is not in the log, the item is removed and the inode it points
1955 * to is unlinked
1956 */
1957static noinline int check_item_in_log(struct btrfs_trans_handle *trans,
1958 struct btrfs_root *root,
1959 struct btrfs_root *log,
1960 struct btrfs_path *path,
1961 struct btrfs_path *log_path,
1962 struct inode *dir,
1963 struct btrfs_key *dir_key)
1964{
1965 int ret;
1966 struct extent_buffer *eb;
1967 int slot;
1968 u32 item_size;
1969 struct btrfs_dir_item *di;
1970 struct btrfs_dir_item *log_di;
1971 int name_len;
1972 unsigned long ptr;
1973 unsigned long ptr_end;
1974 char *name;
1975 struct inode *inode;
1976 struct btrfs_key location;
1977
1978again:
1979 eb = path->nodes[0];
1980 slot = path->slots[0];
1981 item_size = btrfs_item_size_nr(eb, slot);
1982 ptr = btrfs_item_ptr_offset(eb, slot);
1983 ptr_end = ptr + item_size;
1984 while (ptr < ptr_end) {
1985 di = (struct btrfs_dir_item *)ptr;
1986 if (verify_dir_item(root, eb, di)) {
1987 ret = -EIO;
1988 goto out;
1989 }
1990
1991 name_len = btrfs_dir_name_len(eb, di);
1992 name = kmalloc(name_len, GFP_NOFS);
1993 if (!name) {
1994 ret = -ENOMEM;
1995 goto out;
1996 }
1997 read_extent_buffer(eb, name, (unsigned long)(di + 1),
1998 name_len);
1999 log_di = NULL;
2000 if (log && dir_key->type == BTRFS_DIR_ITEM_KEY) {
2001 log_di = btrfs_lookup_dir_item(trans, log, log_path,
2002 dir_key->objectid,
2003 name, name_len, 0);
2004 } else if (log && dir_key->type == BTRFS_DIR_INDEX_KEY) {
2005 log_di = btrfs_lookup_dir_index_item(trans, log,
2006 log_path,
2007 dir_key->objectid,
2008 dir_key->offset,
2009 name, name_len, 0);
2010 }
2011 if (!log_di || (IS_ERR(log_di) && PTR_ERR(log_di) == -ENOENT)) {
2012 btrfs_dir_item_key_to_cpu(eb, di, &location);
2013 btrfs_release_path(path);
2014 btrfs_release_path(log_path);
2015 inode = read_one_inode(root, location.objectid);
2016 if (!inode) {
2017 kfree(name);
2018 return -EIO;
2019 }
2020
2021 ret = link_to_fixup_dir(trans, root,
2022 path, location.objectid);
2023 if (ret) {
2024 kfree(name);
2025 iput(inode);
2026 goto out;
2027 }
2028
2029 inc_nlink(inode);
2030 ret = btrfs_unlink_inode(trans, root, dir, inode,
2031 name, name_len);
2032 if (!ret)
2033 ret = btrfs_run_delayed_items(trans, root);
2034 kfree(name);
2035 iput(inode);
2036 if (ret)
2037 goto out;
2038
2039 /* there might still be more names under this key
2040 * check and repeat if required
2041 */
2042 ret = btrfs_search_slot(NULL, root, dir_key, path,
2043 0, 0);
2044 if (ret == 0)
2045 goto again;
2046 ret = 0;
2047 goto out;
2048 } else if (IS_ERR(log_di)) {
2049 kfree(name);
2050 return PTR_ERR(log_di);
2051 }
2052 btrfs_release_path(log_path);
2053 kfree(name);
2054
2055 ptr = (unsigned long)(di + 1);
2056 ptr += name_len;
2057 }
2058 ret = 0;
2059out:
2060 btrfs_release_path(path);
2061 btrfs_release_path(log_path);
2062 return ret;
2063}
2064
2065static int replay_xattr_deletes(struct btrfs_trans_handle *trans,
2066 struct btrfs_root *root,
2067 struct btrfs_root *log,
2068 struct btrfs_path *path,
2069 const u64 ino)
2070{
2071 struct btrfs_key search_key;
2072 struct btrfs_path *log_path;
2073 int i;
2074 int nritems;
2075 int ret;
2076
2077 log_path = btrfs_alloc_path();
2078 if (!log_path)
2079 return -ENOMEM;
2080
2081 search_key.objectid = ino;
2082 search_key.type = BTRFS_XATTR_ITEM_KEY;
2083 search_key.offset = 0;
2084again:
2085 ret = btrfs_search_slot(NULL, root, &search_key, path, 0, 0);
2086 if (ret < 0)
2087 goto out;
2088process_leaf:
2089 nritems = btrfs_header_nritems(path->nodes[0]);
2090 for (i = path->slots[0]; i < nritems; i++) {
2091 struct btrfs_key key;
2092 struct btrfs_dir_item *di;
2093 struct btrfs_dir_item *log_di;
2094 u32 total_size;
2095 u32 cur;
2096
2097 btrfs_item_key_to_cpu(path->nodes[0], &key, i);
2098 if (key.objectid != ino || key.type != BTRFS_XATTR_ITEM_KEY) {
2099 ret = 0;
2100 goto out;
2101 }
2102
2103 di = btrfs_item_ptr(path->nodes[0], i, struct btrfs_dir_item);
2104 total_size = btrfs_item_size_nr(path->nodes[0], i);
2105 cur = 0;
2106 while (cur < total_size) {
2107 u16 name_len = btrfs_dir_name_len(path->nodes[0], di);
2108 u16 data_len = btrfs_dir_data_len(path->nodes[0], di);
2109 u32 this_len = sizeof(*di) + name_len + data_len;
2110 char *name;
2111
2112 name = kmalloc(name_len, GFP_NOFS);
2113 if (!name) {
2114 ret = -ENOMEM;
2115 goto out;
2116 }
2117 read_extent_buffer(path->nodes[0], name,
2118 (unsigned long)(di + 1), name_len);
2119
2120 log_di = btrfs_lookup_xattr(NULL, log, log_path, ino,
2121 name, name_len, 0);
2122 btrfs_release_path(log_path);
2123 if (!log_di) {
2124 /* Doesn't exist in log tree, so delete it. */
2125 btrfs_release_path(path);
2126 di = btrfs_lookup_xattr(trans, root, path, ino,
2127 name, name_len, -1);
2128 kfree(name);
2129 if (IS_ERR(di)) {
2130 ret = PTR_ERR(di);
2131 goto out;
2132 }
2133 ASSERT(di);
2134 ret = btrfs_delete_one_dir_name(trans, root,
2135 path, di);
2136 if (ret)
2137 goto out;
2138 btrfs_release_path(path);
2139 search_key = key;
2140 goto again;
2141 }
2142 kfree(name);
2143 if (IS_ERR(log_di)) {
2144 ret = PTR_ERR(log_di);
2145 goto out;
2146 }
2147 cur += this_len;
2148 di = (struct btrfs_dir_item *)((char *)di + this_len);
2149 }
2150 }
2151 ret = btrfs_next_leaf(root, path);
2152 if (ret > 0)
2153 ret = 0;
2154 else if (ret == 0)
2155 goto process_leaf;
2156out:
2157 btrfs_free_path(log_path);
2158 btrfs_release_path(path);
2159 return ret;
2160}
2161
2162
2163/*
2164 * deletion replay happens before we copy any new directory items
2165 * out of the log or out of backreferences from inodes. It
2166 * scans the log to find ranges of keys that log is authoritative for,
2167 * and then scans the directory to find items in those ranges that are
2168 * not present in the log.
2169 *
2170 * Anything we don't find in the log is unlinked and removed from the
2171 * directory.
2172 */
2173static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
2174 struct btrfs_root *root,
2175 struct btrfs_root *log,
2176 struct btrfs_path *path,
2177 u64 dirid, int del_all)
2178{
2179 u64 range_start;
2180 u64 range_end;
2181 int key_type = BTRFS_DIR_LOG_ITEM_KEY;
2182 int ret = 0;
2183 struct btrfs_key dir_key;
2184 struct btrfs_key found_key;
2185 struct btrfs_path *log_path;
2186 struct inode *dir;
2187
2188 dir_key.objectid = dirid;
2189 dir_key.type = BTRFS_DIR_ITEM_KEY;
2190 log_path = btrfs_alloc_path();
2191 if (!log_path)
2192 return -ENOMEM;
2193
2194 dir = read_one_inode(root, dirid);
2195 /* it isn't an error if the inode isn't there, that can happen
2196 * because we replay the deletes before we copy in the inode item
2197 * from the log
2198 */
2199 if (!dir) {
2200 btrfs_free_path(log_path);
2201 return 0;
2202 }
2203again:
2204 range_start = 0;
2205 range_end = 0;
2206 while (1) {
2207 if (del_all)
2208 range_end = (u64)-1;
2209 else {
2210 ret = find_dir_range(log, path, dirid, key_type,
2211 &range_start, &range_end);
2212 if (ret != 0)
2213 break;
2214 }
2215
2216 dir_key.offset = range_start;
2217 while (1) {
2218 int nritems;
2219 ret = btrfs_search_slot(NULL, root, &dir_key, path,
2220 0, 0);
2221 if (ret < 0)
2222 goto out;
2223
2224 nritems = btrfs_header_nritems(path->nodes[0]);
2225 if (path->slots[0] >= nritems) {
2226 ret = btrfs_next_leaf(root, path);
2227 if (ret)
2228 break;
2229 }
2230 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
2231 path->slots[0]);
2232 if (found_key.objectid != dirid ||
2233 found_key.type != dir_key.type)
2234 goto next_type;
2235
2236 if (found_key.offset > range_end)
2237 break;
2238
2239 ret = check_item_in_log(trans, root, log, path,
2240 log_path, dir,
2241 &found_key);
2242 if (ret)
2243 goto out;
2244 if (found_key.offset == (u64)-1)
2245 break;
2246 dir_key.offset = found_key.offset + 1;
2247 }
2248 btrfs_release_path(path);
2249 if (range_end == (u64)-1)
2250 break;
2251 range_start = range_end + 1;
2252 }
2253
2254next_type:
2255 ret = 0;
2256 if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
2257 key_type = BTRFS_DIR_LOG_INDEX_KEY;
2258 dir_key.type = BTRFS_DIR_INDEX_KEY;
2259 btrfs_release_path(path);
2260 goto again;
2261 }
2262out:
2263 btrfs_release_path(path);
2264 btrfs_free_path(log_path);
2265 iput(dir);
2266 return ret;
2267}
2268
2269/*
2270 * the process_func used to replay items from the log tree. This
2271 * gets called in two different stages. The first stage just looks
2272 * for inodes and makes sure they are all copied into the subvolume.
2273 *
2274 * The second stage copies all the other item types from the log into
2275 * the subvolume. The two stage approach is slower, but gets rid of
2276 * lots of complexity around inodes referencing other inodes that exist
2277 * only in the log (references come from either directory items or inode
2278 * back refs).
2279 */
2280static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
2281 struct walk_control *wc, u64 gen)
2282{
2283 int nritems;
2284 struct btrfs_path *path;
2285 struct btrfs_root *root = wc->replay_dest;
2286 struct btrfs_key key;
2287 int level;
2288 int i;
2289 int ret;
2290
2291 ret = btrfs_read_buffer(eb, gen);
2292 if (ret)
2293 return ret;
2294
2295 level = btrfs_header_level(eb);
2296
2297 if (level != 0)
2298 return 0;
2299
2300 path = btrfs_alloc_path();
2301 if (!path)
2302 return -ENOMEM;
2303
2304 nritems = btrfs_header_nritems(eb);
2305 for (i = 0; i < nritems; i++) {
2306 btrfs_item_key_to_cpu(eb, &key, i);
2307
2308 /* inode keys are done during the first stage */
2309 if (key.type == BTRFS_INODE_ITEM_KEY &&
2310 wc->stage == LOG_WALK_REPLAY_INODES) {
2311 struct btrfs_inode_item *inode_item;
2312 u32 mode;
2313
2314 inode_item = btrfs_item_ptr(eb, i,
2315 struct btrfs_inode_item);
2316 ret = replay_xattr_deletes(wc->trans, root, log,
2317 path, key.objectid);
2318 if (ret)
2319 break;
2320 mode = btrfs_inode_mode(eb, inode_item);
2321 if (S_ISDIR(mode)) {
2322 ret = replay_dir_deletes(wc->trans,
2323 root, log, path, key.objectid, 0);
2324 if (ret)
2325 break;
2326 }
2327 ret = overwrite_item(wc->trans, root, path,
2328 eb, i, &key);
2329 if (ret)
2330 break;
2331
2332 /* for regular files, make sure corresponding
2333 * orhpan item exist. extents past the new EOF
2334 * will be truncated later by orphan cleanup.
2335 */
2336 if (S_ISREG(mode)) {
2337 ret = insert_orphan_item(wc->trans, root,
2338 key.objectid);
2339 if (ret)
2340 break;
2341 }
2342
2343 ret = link_to_fixup_dir(wc->trans, root,
2344 path, key.objectid);
2345 if (ret)
2346 break;
2347 }
2348
2349 if (key.type == BTRFS_DIR_INDEX_KEY &&
2350 wc->stage == LOG_WALK_REPLAY_DIR_INDEX) {
2351 ret = replay_one_dir_item(wc->trans, root, path,
2352 eb, i, &key);
2353 if (ret)
2354 break;
2355 }
2356
2357 if (wc->stage < LOG_WALK_REPLAY_ALL)
2358 continue;
2359
2360 /* these keys are simply copied */
2361 if (key.type == BTRFS_XATTR_ITEM_KEY) {
2362 ret = overwrite_item(wc->trans, root, path,
2363 eb, i, &key);
2364 if (ret)
2365 break;
2366 } else if (key.type == BTRFS_INODE_REF_KEY ||
2367 key.type == BTRFS_INODE_EXTREF_KEY) {
2368 ret = add_inode_ref(wc->trans, root, log, path,
2369 eb, i, &key);
2370 if (ret && ret != -ENOENT)
2371 break;
2372 ret = 0;
2373 } else if (key.type == BTRFS_EXTENT_DATA_KEY) {
2374 ret = replay_one_extent(wc->trans, root, path,
2375 eb, i, &key);
2376 if (ret)
2377 break;
2378 } else if (key.type == BTRFS_DIR_ITEM_KEY) {
2379 ret = replay_one_dir_item(wc->trans, root, path,
2380 eb, i, &key);
2381 if (ret)
2382 break;
2383 }
2384 }
2385 btrfs_free_path(path);
2386 return ret;
2387}
2388
2389static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
2390 struct btrfs_root *root,
2391 struct btrfs_path *path, int *level,
2392 struct walk_control *wc)
2393{
2394 u64 root_owner;
2395 u64 bytenr;
2396 u64 ptr_gen;
2397 struct extent_buffer *next;
2398 struct extent_buffer *cur;
2399 struct extent_buffer *parent;
2400 u32 blocksize;
2401 int ret = 0;
2402
2403 WARN_ON(*level < 0);
2404 WARN_ON(*level >= BTRFS_MAX_LEVEL);
2405
2406 while (*level > 0) {
2407 WARN_ON(*level < 0);
2408 WARN_ON(*level >= BTRFS_MAX_LEVEL);
2409 cur = path->nodes[*level];
2410
2411 WARN_ON(btrfs_header_level(cur) != *level);
2412
2413 if (path->slots[*level] >=
2414 btrfs_header_nritems(cur))
2415 break;
2416
2417 bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
2418 ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
2419 blocksize = root->nodesize;
2420
2421 parent = path->nodes[*level];
2422 root_owner = btrfs_header_owner(parent);
2423
2424 next = btrfs_find_create_tree_block(root, bytenr);
2425 if (!next)
2426 return -ENOMEM;
2427
2428 if (*level == 1) {
2429 ret = wc->process_func(root, next, wc, ptr_gen);
2430 if (ret) {
2431 free_extent_buffer(next);
2432 return ret;
2433 }
2434
2435 path->slots[*level]++;
2436 if (wc->free) {
2437 ret = btrfs_read_buffer(next, ptr_gen);
2438 if (ret) {
2439 free_extent_buffer(next);
2440 return ret;
2441 }
2442
2443 if (trans) {
2444 btrfs_tree_lock(next);
2445 btrfs_set_lock_blocking(next);
2446 clean_tree_block(trans, root->fs_info,
2447 next);
2448 btrfs_wait_tree_block_writeback(next);
2449 btrfs_tree_unlock(next);
2450 }
2451
2452 WARN_ON(root_owner !=
2453 BTRFS_TREE_LOG_OBJECTID);
2454 ret = btrfs_free_and_pin_reserved_extent(root,
2455 bytenr, blocksize);
2456 if (ret) {
2457 free_extent_buffer(next);
2458 return ret;
2459 }
2460 }
2461 free_extent_buffer(next);
2462 continue;
2463 }
2464 ret = btrfs_read_buffer(next, ptr_gen);
2465 if (ret) {
2466 free_extent_buffer(next);
2467 return ret;
2468 }
2469
2470 WARN_ON(*level <= 0);
2471 if (path->nodes[*level-1])
2472 free_extent_buffer(path->nodes[*level-1]);
2473 path->nodes[*level-1] = next;
2474 *level = btrfs_header_level(next);
2475 path->slots[*level] = 0;
2476 cond_resched();
2477 }
2478 WARN_ON(*level < 0);
2479 WARN_ON(*level >= BTRFS_MAX_LEVEL);
2480
2481 path->slots[*level] = btrfs_header_nritems(path->nodes[*level]);
2482
2483 cond_resched();
2484 return 0;
2485}
2486
2487static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
2488 struct btrfs_root *root,
2489 struct btrfs_path *path, int *level,
2490 struct walk_control *wc)
2491{
2492 u64 root_owner;
2493 int i;
2494 int slot;
2495 int ret;
2496
2497 for (i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
2498 slot = path->slots[i];
2499 if (slot + 1 < btrfs_header_nritems(path->nodes[i])) {
2500 path->slots[i]++;
2501 *level = i;
2502 WARN_ON(*level == 0);
2503 return 0;
2504 } else {
2505 struct extent_buffer *parent;
2506 if (path->nodes[*level] == root->node)
2507 parent = path->nodes[*level];
2508 else
2509 parent = path->nodes[*level + 1];
2510
2511 root_owner = btrfs_header_owner(parent);
2512 ret = wc->process_func(root, path->nodes[*level], wc,
2513 btrfs_header_generation(path->nodes[*level]));
2514 if (ret)
2515 return ret;
2516
2517 if (wc->free) {
2518 struct extent_buffer *next;
2519
2520 next = path->nodes[*level];
2521
2522 if (trans) {
2523 btrfs_tree_lock(next);
2524 btrfs_set_lock_blocking(next);
2525 clean_tree_block(trans, root->fs_info,
2526 next);
2527 btrfs_wait_tree_block_writeback(next);
2528 btrfs_tree_unlock(next);
2529 }
2530
2531 WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
2532 ret = btrfs_free_and_pin_reserved_extent(root,
2533 path->nodes[*level]->start,
2534 path->nodes[*level]->len);
2535 if (ret)
2536 return ret;
2537 }
2538 free_extent_buffer(path->nodes[*level]);
2539 path->nodes[*level] = NULL;
2540 *level = i + 1;
2541 }
2542 }
2543 return 1;
2544}
2545
2546/*
2547 * drop the reference count on the tree rooted at 'snap'. This traverses
2548 * the tree freeing any blocks that have a ref count of zero after being
2549 * decremented.
2550 */
2551static int walk_log_tree(struct btrfs_trans_handle *trans,
2552 struct btrfs_root *log, struct walk_control *wc)
2553{
2554 int ret = 0;
2555 int wret;
2556 int level;
2557 struct btrfs_path *path;
2558 int orig_level;
2559
2560 path = btrfs_alloc_path();
2561 if (!path)
2562 return -ENOMEM;
2563
2564 level = btrfs_header_level(log->node);
2565 orig_level = level;
2566 path->nodes[level] = log->node;
2567 extent_buffer_get(log->node);
2568 path->slots[level] = 0;
2569
2570 while (1) {
2571 wret = walk_down_log_tree(trans, log, path, &level, wc);
2572 if (wret > 0)
2573 break;
2574 if (wret < 0) {
2575 ret = wret;
2576 goto out;
2577 }
2578
2579 wret = walk_up_log_tree(trans, log, path, &level, wc);
2580 if (wret > 0)
2581 break;
2582 if (wret < 0) {
2583 ret = wret;
2584 goto out;
2585 }
2586 }
2587
2588 /* was the root node processed? if not, catch it here */
2589 if (path->nodes[orig_level]) {
2590 ret = wc->process_func(log, path->nodes[orig_level], wc,
2591 btrfs_header_generation(path->nodes[orig_level]));
2592 if (ret)
2593 goto out;
2594 if (wc->free) {
2595 struct extent_buffer *next;
2596
2597 next = path->nodes[orig_level];
2598
2599 if (trans) {
2600 btrfs_tree_lock(next);
2601 btrfs_set_lock_blocking(next);
2602 clean_tree_block(trans, log->fs_info, next);
2603 btrfs_wait_tree_block_writeback(next);
2604 btrfs_tree_unlock(next);
2605 }
2606
2607 WARN_ON(log->root_key.objectid !=
2608 BTRFS_TREE_LOG_OBJECTID);
2609 ret = btrfs_free_and_pin_reserved_extent(log, next->start,
2610 next->len);
2611 if (ret)
2612 goto out;
2613 }
2614 }
2615
2616out:
2617 btrfs_free_path(path);
2618 return ret;
2619}
2620
2621/*
2622 * helper function to update the item for a given subvolumes log root
2623 * in the tree of log roots
2624 */
2625static int update_log_root(struct btrfs_trans_handle *trans,
2626 struct btrfs_root *log)
2627{
2628 int ret;
2629
2630 if (log->log_transid == 1) {
2631 /* insert root item on the first sync */
2632 ret = btrfs_insert_root(trans, log->fs_info->log_root_tree,
2633 &log->root_key, &log->root_item);
2634 } else {
2635 ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
2636 &log->root_key, &log->root_item);
2637 }
2638 return ret;
2639}
2640
2641static void wait_log_commit(struct btrfs_root *root, int transid)
2642{
2643 DEFINE_WAIT(wait);
2644 int index = transid % 2;
2645
2646 /*
2647 * we only allow two pending log transactions at a time,
2648 * so we know that if ours is more than 2 older than the
2649 * current transaction, we're done
2650 */
2651 do {
2652 prepare_to_wait(&root->log_commit_wait[index],
2653 &wait, TASK_UNINTERRUPTIBLE);
2654 mutex_unlock(&root->log_mutex);
2655
2656 if (root->log_transid_committed < transid &&
2657 atomic_read(&root->log_commit[index]))
2658 schedule();
2659
2660 finish_wait(&root->log_commit_wait[index], &wait);
2661 mutex_lock(&root->log_mutex);
2662 } while (root->log_transid_committed < transid &&
2663 atomic_read(&root->log_commit[index]));
2664}
2665
2666static void wait_for_writer(struct btrfs_root *root)
2667{
2668 DEFINE_WAIT(wait);
2669
2670 while (atomic_read(&root->log_writers)) {
2671 prepare_to_wait(&root->log_writer_wait,
2672 &wait, TASK_UNINTERRUPTIBLE);
2673 mutex_unlock(&root->log_mutex);
2674 if (atomic_read(&root->log_writers))
2675 schedule();
2676 finish_wait(&root->log_writer_wait, &wait);
2677 mutex_lock(&root->log_mutex);
2678 }
2679}
2680
2681static inline void btrfs_remove_log_ctx(struct btrfs_root *root,
2682 struct btrfs_log_ctx *ctx)
2683{
2684 if (!ctx)
2685 return;
2686
2687 mutex_lock(&root->log_mutex);
2688 list_del_init(&ctx->list);
2689 mutex_unlock(&root->log_mutex);
2690}
2691
2692/*
2693 * Invoked in log mutex context, or be sure there is no other task which
2694 * can access the list.
2695 */
2696static inline void btrfs_remove_all_log_ctxs(struct btrfs_root *root,
2697 int index, int error)
2698{
2699 struct btrfs_log_ctx *ctx;
2700
2701 if (!error) {
2702 INIT_LIST_HEAD(&root->log_ctxs[index]);
2703 return;
2704 }
2705
2706 list_for_each_entry(ctx, &root->log_ctxs[index], list)
2707 ctx->log_ret = error;
2708
2709 INIT_LIST_HEAD(&root->log_ctxs[index]);
2710}
2711
2712/*
2713 * btrfs_sync_log does sends a given tree log down to the disk and
2714 * updates the super blocks to record it. When this call is done,
2715 * you know that any inodes previously logged are safely on disk only
2716 * if it returns 0.
2717 *
2718 * Any other return value means you need to call btrfs_commit_transaction.
2719 * Some of the edge cases for fsyncing directories that have had unlinks
2720 * or renames done in the past mean that sometimes the only safe
2721 * fsync is to commit the whole FS. When btrfs_sync_log returns -EAGAIN,
2722 * that has happened.
2723 */
2724int btrfs_sync_log(struct btrfs_trans_handle *trans,
2725 struct btrfs_root *root, struct btrfs_log_ctx *ctx)
2726{
2727 int index1;
2728 int index2;
2729 int mark;
2730 int ret;
2731 struct btrfs_root *log = root->log_root;
2732 struct btrfs_root *log_root_tree = root->fs_info->log_root_tree;
2733 int log_transid = 0;
2734 struct btrfs_log_ctx root_log_ctx;
2735 struct blk_plug plug;
2736
2737 mutex_lock(&root->log_mutex);
2738 log_transid = ctx->log_transid;
2739 if (root->log_transid_committed >= log_transid) {
2740 mutex_unlock(&root->log_mutex);
2741 return ctx->log_ret;
2742 }
2743
2744 index1 = log_transid % 2;
2745 if (atomic_read(&root->log_commit[index1])) {
2746 wait_log_commit(root, log_transid);
2747 mutex_unlock(&root->log_mutex);
2748 return ctx->log_ret;
2749 }
2750 ASSERT(log_transid == root->log_transid);
2751 atomic_set(&root->log_commit[index1], 1);
2752
2753 /* wait for previous tree log sync to complete */
2754 if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
2755 wait_log_commit(root, log_transid - 1);
2756
2757 while (1) {
2758 int batch = atomic_read(&root->log_batch);
2759 /* when we're on an ssd, just kick the log commit out */
2760 if (!btrfs_test_opt(root, SSD) &&
2761 test_bit(BTRFS_ROOT_MULTI_LOG_TASKS, &root->state)) {
2762 mutex_unlock(&root->log_mutex);
2763 schedule_timeout_uninterruptible(1);
2764 mutex_lock(&root->log_mutex);
2765 }
2766 wait_for_writer(root);
2767 if (batch == atomic_read(&root->log_batch))
2768 break;
2769 }
2770
2771 /* bail out if we need to do a full commit */
2772 if (btrfs_need_log_full_commit(root->fs_info, trans)) {
2773 ret = -EAGAIN;
2774 btrfs_free_logged_extents(log, log_transid);
2775 mutex_unlock(&root->log_mutex);
2776 goto out;
2777 }
2778
2779 if (log_transid % 2 == 0)
2780 mark = EXTENT_DIRTY;
2781 else
2782 mark = EXTENT_NEW;
2783
2784 /* we start IO on all the marked extents here, but we don't actually
2785 * wait for them until later.
2786 */
2787 blk_start_plug(&plug);
2788 ret = btrfs_write_marked_extents(log, &log->dirty_log_pages, mark);
2789 if (ret) {
2790 blk_finish_plug(&plug);
2791 btrfs_abort_transaction(trans, root, ret);
2792 btrfs_free_logged_extents(log, log_transid);
2793 btrfs_set_log_full_commit(root->fs_info, trans);
2794 mutex_unlock(&root->log_mutex);
2795 goto out;
2796 }
2797
2798 btrfs_set_root_node(&log->root_item, log->node);
2799
2800 root->log_transid++;
2801 log->log_transid = root->log_transid;
2802 root->log_start_pid = 0;
2803 /*
2804 * IO has been started, blocks of the log tree have WRITTEN flag set
2805 * in their headers. new modifications of the log will be written to
2806 * new positions. so it's safe to allow log writers to go in.
2807 */
2808 mutex_unlock(&root->log_mutex);
2809
2810 btrfs_init_log_ctx(&root_log_ctx);
2811
2812 mutex_lock(&log_root_tree->log_mutex);
2813 atomic_inc(&log_root_tree->log_batch);
2814 atomic_inc(&log_root_tree->log_writers);
2815
2816 index2 = log_root_tree->log_transid % 2;
2817 list_add_tail(&root_log_ctx.list, &log_root_tree->log_ctxs[index2]);
2818 root_log_ctx.log_transid = log_root_tree->log_transid;
2819
2820 mutex_unlock(&log_root_tree->log_mutex);
2821
2822 ret = update_log_root(trans, log);
2823
2824 mutex_lock(&log_root_tree->log_mutex);
2825 if (atomic_dec_and_test(&log_root_tree->log_writers)) {
2826 /*
2827 * Implicit memory barrier after atomic_dec_and_test
2828 */
2829 if (waitqueue_active(&log_root_tree->log_writer_wait))
2830 wake_up(&log_root_tree->log_writer_wait);
2831 }
2832
2833 if (ret) {
2834 if (!list_empty(&root_log_ctx.list))
2835 list_del_init(&root_log_ctx.list);
2836
2837 blk_finish_plug(&plug);
2838 btrfs_set_log_full_commit(root->fs_info, trans);
2839
2840 if (ret != -ENOSPC) {
2841 btrfs_abort_transaction(trans, root, ret);
2842 mutex_unlock(&log_root_tree->log_mutex);
2843 goto out;
2844 }
2845 btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2846 btrfs_free_logged_extents(log, log_transid);
2847 mutex_unlock(&log_root_tree->log_mutex);
2848 ret = -EAGAIN;
2849 goto out;
2850 }
2851
2852 if (log_root_tree->log_transid_committed >= root_log_ctx.log_transid) {
2853 blk_finish_plug(&plug);
2854 mutex_unlock(&log_root_tree->log_mutex);
2855 ret = root_log_ctx.log_ret;
2856 goto out;
2857 }
2858
2859 index2 = root_log_ctx.log_transid % 2;
2860 if (atomic_read(&log_root_tree->log_commit[index2])) {
2861 blk_finish_plug(&plug);
2862 ret = btrfs_wait_marked_extents(log, &log->dirty_log_pages,
2863 mark);
2864 btrfs_wait_logged_extents(trans, log, log_transid);
2865 wait_log_commit(log_root_tree,
2866 root_log_ctx.log_transid);
2867 mutex_unlock(&log_root_tree->log_mutex);
2868 if (!ret)
2869 ret = root_log_ctx.log_ret;
2870 goto out;
2871 }
2872 ASSERT(root_log_ctx.log_transid == log_root_tree->log_transid);
2873 atomic_set(&log_root_tree->log_commit[index2], 1);
2874
2875 if (atomic_read(&log_root_tree->log_commit[(index2 + 1) % 2])) {
2876 wait_log_commit(log_root_tree,
2877 root_log_ctx.log_transid - 1);
2878 }
2879
2880 wait_for_writer(log_root_tree);
2881
2882 /*
2883 * now that we've moved on to the tree of log tree roots,
2884 * check the full commit flag again
2885 */
2886 if (btrfs_need_log_full_commit(root->fs_info, trans)) {
2887 blk_finish_plug(&plug);
2888 btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2889 btrfs_free_logged_extents(log, log_transid);
2890 mutex_unlock(&log_root_tree->log_mutex);
2891 ret = -EAGAIN;
2892 goto out_wake_log_root;
2893 }
2894
2895 ret = btrfs_write_marked_extents(log_root_tree,
2896 &log_root_tree->dirty_log_pages,
2897 EXTENT_DIRTY | EXTENT_NEW);
2898 blk_finish_plug(&plug);
2899 if (ret) {
2900 btrfs_set_log_full_commit(root->fs_info, trans);
2901 btrfs_abort_transaction(trans, root, ret);
2902 btrfs_free_logged_extents(log, log_transid);
2903 mutex_unlock(&log_root_tree->log_mutex);
2904 goto out_wake_log_root;
2905 }
2906 ret = btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2907 if (!ret)
2908 ret = btrfs_wait_marked_extents(log_root_tree,
2909 &log_root_tree->dirty_log_pages,
2910 EXTENT_NEW | EXTENT_DIRTY);
2911 if (ret) {
2912 btrfs_set_log_full_commit(root->fs_info, trans);
2913 btrfs_free_logged_extents(log, log_transid);
2914 mutex_unlock(&log_root_tree->log_mutex);
2915 goto out_wake_log_root;
2916 }
2917 btrfs_wait_logged_extents(trans, log, log_transid);
2918
2919 btrfs_set_super_log_root(root->fs_info->super_for_commit,
2920 log_root_tree->node->start);
2921 btrfs_set_super_log_root_level(root->fs_info->super_for_commit,
2922 btrfs_header_level(log_root_tree->node));
2923
2924 log_root_tree->log_transid++;
2925 mutex_unlock(&log_root_tree->log_mutex);
2926
2927 /*
2928 * nobody else is going to jump in and write the the ctree
2929 * super here because the log_commit atomic below is protecting
2930 * us. We must be called with a transaction handle pinning
2931 * the running transaction open, so a full commit can't hop
2932 * in and cause problems either.
2933 */
2934 ret = write_ctree_super(trans, root->fs_info->tree_root, 1);
2935 if (ret) {
2936 btrfs_set_log_full_commit(root->fs_info, trans);
2937 btrfs_abort_transaction(trans, root, ret);
2938 goto out_wake_log_root;
2939 }
2940
2941 mutex_lock(&root->log_mutex);
2942 if (root->last_log_commit < log_transid)
2943 root->last_log_commit = log_transid;
2944 mutex_unlock(&root->log_mutex);
2945
2946out_wake_log_root:
2947 /*
2948 * We needn't get log_mutex here because we are sure all
2949 * the other tasks are blocked.
2950 */
2951 btrfs_remove_all_log_ctxs(log_root_tree, index2, ret);
2952
2953 mutex_lock(&log_root_tree->log_mutex);
2954 log_root_tree->log_transid_committed++;
2955 atomic_set(&log_root_tree->log_commit[index2], 0);
2956 mutex_unlock(&log_root_tree->log_mutex);
2957
2958 /*
2959 * The barrier before waitqueue_active is implied by mutex_unlock
2960 */
2961 if (waitqueue_active(&log_root_tree->log_commit_wait[index2]))
2962 wake_up(&log_root_tree->log_commit_wait[index2]);
2963out:
2964 /* See above. */
2965 btrfs_remove_all_log_ctxs(root, index1, ret);
2966
2967 mutex_lock(&root->log_mutex);
2968 root->log_transid_committed++;
2969 atomic_set(&root->log_commit[index1], 0);
2970 mutex_unlock(&root->log_mutex);
2971
2972 /*
2973 * The barrier before waitqueue_active is implied by mutex_unlock
2974 */
2975 if (waitqueue_active(&root->log_commit_wait[index1]))
2976 wake_up(&root->log_commit_wait[index1]);
2977 return ret;
2978}
2979
2980static void free_log_tree(struct btrfs_trans_handle *trans,
2981 struct btrfs_root *log)
2982{
2983 int ret;
2984 u64 start;
2985 u64 end;
2986 struct walk_control wc = {
2987 .free = 1,
2988 .process_func = process_one_buffer
2989 };
2990
2991 ret = walk_log_tree(trans, log, &wc);
2992 /* I don't think this can happen but just in case */
2993 if (ret)
2994 btrfs_abort_transaction(trans, log, ret);
2995
2996 while (1) {
2997 ret = find_first_extent_bit(&log->dirty_log_pages,
2998 0, &start, &end, EXTENT_DIRTY | EXTENT_NEW,
2999 NULL);
3000 if (ret)
3001 break;
3002
3003 clear_extent_bits(&log->dirty_log_pages, start, end,
3004 EXTENT_DIRTY | EXTENT_NEW, GFP_NOFS);
3005 }
3006
3007 /*
3008 * We may have short-circuited the log tree with the full commit logic
3009 * and left ordered extents on our list, so clear these out to keep us
3010 * from leaking inodes and memory.
3011 */
3012 btrfs_free_logged_extents(log, 0);
3013 btrfs_free_logged_extents(log, 1);
3014
3015 free_extent_buffer(log->node);
3016 kfree(log);
3017}
3018
3019/*
3020 * free all the extents used by the tree log. This should be called
3021 * at commit time of the full transaction
3022 */
3023int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
3024{
3025 if (root->log_root) {
3026 free_log_tree(trans, root->log_root);
3027 root->log_root = NULL;
3028 }
3029 return 0;
3030}
3031
3032int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
3033 struct btrfs_fs_info *fs_info)
3034{
3035 if (fs_info->log_root_tree) {
3036 free_log_tree(trans, fs_info->log_root_tree);
3037 fs_info->log_root_tree = NULL;
3038 }
3039 return 0;
3040}
3041
3042/*
3043 * If both a file and directory are logged, and unlinks or renames are
3044 * mixed in, we have a few interesting corners:
3045 *
3046 * create file X in dir Y
3047 * link file X to X.link in dir Y
3048 * fsync file X
3049 * unlink file X but leave X.link
3050 * fsync dir Y
3051 *
3052 * After a crash we would expect only X.link to exist. But file X
3053 * didn't get fsync'd again so the log has back refs for X and X.link.
3054 *
3055 * We solve this by removing directory entries and inode backrefs from the
3056 * log when a file that was logged in the current transaction is
3057 * unlinked. Any later fsync will include the updated log entries, and
3058 * we'll be able to reconstruct the proper directory items from backrefs.
3059 *
3060 * This optimizations allows us to avoid relogging the entire inode
3061 * or the entire directory.
3062 */
3063int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
3064 struct btrfs_root *root,
3065 const char *name, int name_len,
3066 struct inode *dir, u64 index)
3067{
3068 struct btrfs_root *log;
3069 struct btrfs_dir_item *di;
3070 struct btrfs_path *path;
3071 int ret;
3072 int err = 0;
3073 int bytes_del = 0;
3074 u64 dir_ino = btrfs_ino(dir);
3075
3076 if (BTRFS_I(dir)->logged_trans < trans->transid)
3077 return 0;
3078
3079 ret = join_running_log_trans(root);
3080 if (ret)
3081 return 0;
3082
3083 mutex_lock(&BTRFS_I(dir)->log_mutex);
3084
3085 log = root->log_root;
3086 path = btrfs_alloc_path();
3087 if (!path) {
3088 err = -ENOMEM;
3089 goto out_unlock;
3090 }
3091
3092 di = btrfs_lookup_dir_item(trans, log, path, dir_ino,
3093 name, name_len, -1);
3094 if (IS_ERR(di)) {
3095 err = PTR_ERR(di);
3096 goto fail;
3097 }
3098 if (di) {
3099 ret = btrfs_delete_one_dir_name(trans, log, path, di);
3100 bytes_del += name_len;
3101 if (ret) {
3102 err = ret;
3103 goto fail;
3104 }
3105 }
3106 btrfs_release_path(path);
3107 di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino,
3108 index, name, name_len, -1);
3109 if (IS_ERR(di)) {
3110 err = PTR_ERR(di);
3111 goto fail;
3112 }
3113 if (di) {
3114 ret = btrfs_delete_one_dir_name(trans, log, path, di);
3115 bytes_del += name_len;
3116 if (ret) {
3117 err = ret;
3118 goto fail;
3119 }
3120 }
3121
3122 /* update the directory size in the log to reflect the names
3123 * we have removed
3124 */
3125 if (bytes_del) {
3126 struct btrfs_key key;
3127
3128 key.objectid = dir_ino;
3129 key.offset = 0;
3130 key.type = BTRFS_INODE_ITEM_KEY;
3131 btrfs_release_path(path);
3132
3133 ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
3134 if (ret < 0) {
3135 err = ret;
3136 goto fail;
3137 }
3138 if (ret == 0) {
3139 struct btrfs_inode_item *item;
3140 u64 i_size;
3141
3142 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
3143 struct btrfs_inode_item);
3144 i_size = btrfs_inode_size(path->nodes[0], item);
3145 if (i_size > bytes_del)
3146 i_size -= bytes_del;
3147 else
3148 i_size = 0;
3149 btrfs_set_inode_size(path->nodes[0], item, i_size);
3150 btrfs_mark_buffer_dirty(path->nodes[0]);
3151 } else
3152 ret = 0;
3153 btrfs_release_path(path);
3154 }
3155fail:
3156 btrfs_free_path(path);
3157out_unlock:
3158 mutex_unlock(&BTRFS_I(dir)->log_mutex);
3159 if (ret == -ENOSPC) {
3160 btrfs_set_log_full_commit(root->fs_info, trans);
3161 ret = 0;
3162 } else if (ret < 0)
3163 btrfs_abort_transaction(trans, root, ret);
3164
3165 btrfs_end_log_trans(root);
3166
3167 return err;
3168}
3169
3170/* see comments for btrfs_del_dir_entries_in_log */
3171int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
3172 struct btrfs_root *root,
3173 const char *name, int name_len,
3174 struct inode *inode, u64 dirid)
3175{
3176 struct btrfs_root *log;
3177 u64 index;
3178 int ret;
3179
3180 if (BTRFS_I(inode)->logged_trans < trans->transid)
3181 return 0;
3182
3183 ret = join_running_log_trans(root);
3184 if (ret)
3185 return 0;
3186 log = root->log_root;
3187 mutex_lock(&BTRFS_I(inode)->log_mutex);
3188
3189 ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode),
3190 dirid, &index);
3191 mutex_unlock(&BTRFS_I(inode)->log_mutex);
3192 if (ret == -ENOSPC) {
3193 btrfs_set_log_full_commit(root->fs_info, trans);
3194 ret = 0;
3195 } else if (ret < 0 && ret != -ENOENT)
3196 btrfs_abort_transaction(trans, root, ret);
3197 btrfs_end_log_trans(root);
3198
3199 return ret;
3200}
3201
3202/*
3203 * creates a range item in the log for 'dirid'. first_offset and
3204 * last_offset tell us which parts of the key space the log should
3205 * be considered authoritative for.
3206 */
3207static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
3208 struct btrfs_root *log,
3209 struct btrfs_path *path,
3210 int key_type, u64 dirid,
3211 u64 first_offset, u64 last_offset)
3212{
3213 int ret;
3214 struct btrfs_key key;
3215 struct btrfs_dir_log_item *item;
3216
3217 key.objectid = dirid;
3218 key.offset = first_offset;
3219 if (key_type == BTRFS_DIR_ITEM_KEY)
3220 key.type = BTRFS_DIR_LOG_ITEM_KEY;
3221 else
3222 key.type = BTRFS_DIR_LOG_INDEX_KEY;
3223 ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
3224 if (ret)
3225 return ret;
3226
3227 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
3228 struct btrfs_dir_log_item);
3229 btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
3230 btrfs_mark_buffer_dirty(path->nodes[0]);
3231 btrfs_release_path(path);
3232 return 0;
3233}
3234
3235/*
3236 * log all the items included in the current transaction for a given
3237 * directory. This also creates the range items in the log tree required
3238 * to replay anything deleted before the fsync
3239 */
3240static noinline int log_dir_items(struct btrfs_trans_handle *trans,
3241 struct btrfs_root *root, struct inode *inode,
3242 struct btrfs_path *path,
3243 struct btrfs_path *dst_path, int key_type,
3244 struct btrfs_log_ctx *ctx,
3245 u64 min_offset, u64 *last_offset_ret)
3246{
3247 struct btrfs_key min_key;
3248 struct btrfs_root *log = root->log_root;
3249 struct extent_buffer *src;
3250 int err = 0;
3251 int ret;
3252 int i;
3253 int nritems;
3254 u64 first_offset = min_offset;
3255 u64 last_offset = (u64)-1;
3256 u64 ino = btrfs_ino(inode);
3257
3258 log = root->log_root;
3259
3260 min_key.objectid = ino;
3261 min_key.type = key_type;
3262 min_key.offset = min_offset;
3263
3264 ret = btrfs_search_forward(root, &min_key, path, trans->transid);
3265
3266 /*
3267 * we didn't find anything from this transaction, see if there
3268 * is anything at all
3269 */
3270 if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) {
3271 min_key.objectid = ino;
3272 min_key.type = key_type;
3273 min_key.offset = (u64)-1;
3274 btrfs_release_path(path);
3275 ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
3276 if (ret < 0) {
3277 btrfs_release_path(path);
3278 return ret;
3279 }
3280 ret = btrfs_previous_item(root, path, ino, key_type);
3281
3282 /* if ret == 0 there are items for this type,
3283 * create a range to tell us the last key of this type.
3284 * otherwise, there are no items in this directory after
3285 * *min_offset, and we create a range to indicate that.
3286 */
3287 if (ret == 0) {
3288 struct btrfs_key tmp;
3289 btrfs_item_key_to_cpu(path->nodes[0], &tmp,
3290 path->slots[0]);
3291 if (key_type == tmp.type)
3292 first_offset = max(min_offset, tmp.offset) + 1;
3293 }
3294 goto done;
3295 }
3296
3297 /* go backward to find any previous key */
3298 ret = btrfs_previous_item(root, path, ino, key_type);
3299 if (ret == 0) {
3300 struct btrfs_key tmp;
3301 btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
3302 if (key_type == tmp.type) {
3303 first_offset = tmp.offset;
3304 ret = overwrite_item(trans, log, dst_path,
3305 path->nodes[0], path->slots[0],
3306 &tmp);
3307 if (ret) {
3308 err = ret;
3309 goto done;
3310 }
3311 }
3312 }
3313 btrfs_release_path(path);
3314
3315 /* find the first key from this transaction again */
3316 ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
3317 if (WARN_ON(ret != 0))
3318 goto done;
3319
3320 /*
3321 * we have a block from this transaction, log every item in it
3322 * from our directory
3323 */
3324 while (1) {
3325 struct btrfs_key tmp;
3326 src = path->nodes[0];
3327 nritems = btrfs_header_nritems(src);
3328 for (i = path->slots[0]; i < nritems; i++) {
3329 struct btrfs_dir_item *di;
3330
3331 btrfs_item_key_to_cpu(src, &min_key, i);
3332
3333 if (min_key.objectid != ino || min_key.type != key_type)
3334 goto done;
3335 ret = overwrite_item(trans, log, dst_path, src, i,
3336 &min_key);
3337 if (ret) {
3338 err = ret;
3339 goto done;
3340 }
3341
3342 /*
3343 * We must make sure that when we log a directory entry,
3344 * the corresponding inode, after log replay, has a
3345 * matching link count. For example:
3346 *
3347 * touch foo
3348 * mkdir mydir
3349 * sync
3350 * ln foo mydir/bar
3351 * xfs_io -c "fsync" mydir
3352 * <crash>
3353 * <mount fs and log replay>
3354 *
3355 * Would result in a fsync log that when replayed, our
3356 * file inode would have a link count of 1, but we get
3357 * two directory entries pointing to the same inode.
3358 * After removing one of the names, it would not be
3359 * possible to remove the other name, which resulted
3360 * always in stale file handle errors, and would not
3361 * be possible to rmdir the parent directory, since
3362 * its i_size could never decrement to the value
3363 * BTRFS_EMPTY_DIR_SIZE, resulting in -ENOTEMPTY errors.
3364 */
3365 di = btrfs_item_ptr(src, i, struct btrfs_dir_item);
3366 btrfs_dir_item_key_to_cpu(src, di, &tmp);
3367 if (ctx &&
3368 (btrfs_dir_transid(src, di) == trans->transid ||
3369 btrfs_dir_type(src, di) == BTRFS_FT_DIR) &&
3370 tmp.type != BTRFS_ROOT_ITEM_KEY)
3371 ctx->log_new_dentries = true;
3372 }
3373 path->slots[0] = nritems;
3374
3375 /*
3376 * look ahead to the next item and see if it is also
3377 * from this directory and from this transaction
3378 */
3379 ret = btrfs_next_leaf(root, path);
3380 if (ret == 1) {
3381 last_offset = (u64)-1;
3382 goto done;
3383 }
3384 btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
3385 if (tmp.objectid != ino || tmp.type != key_type) {
3386 last_offset = (u64)-1;
3387 goto done;
3388 }
3389 if (btrfs_header_generation(path->nodes[0]) != trans->transid) {
3390 ret = overwrite_item(trans, log, dst_path,
3391 path->nodes[0], path->slots[0],
3392 &tmp);
3393 if (ret)
3394 err = ret;
3395 else
3396 last_offset = tmp.offset;
3397 goto done;
3398 }
3399 }
3400done:
3401 btrfs_release_path(path);
3402 btrfs_release_path(dst_path);
3403
3404 if (err == 0) {
3405 *last_offset_ret = last_offset;
3406 /*
3407 * insert the log range keys to indicate where the log
3408 * is valid
3409 */
3410 ret = insert_dir_log_key(trans, log, path, key_type,
3411 ino, first_offset, last_offset);
3412 if (ret)
3413 err = ret;
3414 }
3415 return err;
3416}
3417
3418/*
3419 * logging directories is very similar to logging inodes, We find all the items
3420 * from the current transaction and write them to the log.
3421 *
3422 * The recovery code scans the directory in the subvolume, and if it finds a
3423 * key in the range logged that is not present in the log tree, then it means
3424 * that dir entry was unlinked during the transaction.
3425 *
3426 * In order for that scan to work, we must include one key smaller than
3427 * the smallest logged by this transaction and one key larger than the largest
3428 * key logged by this transaction.
3429 */
3430static noinline int log_directory_changes(struct btrfs_trans_handle *trans,
3431 struct btrfs_root *root, struct inode *inode,
3432 struct btrfs_path *path,
3433 struct btrfs_path *dst_path,
3434 struct btrfs_log_ctx *ctx)
3435{
3436 u64 min_key;
3437 u64 max_key;
3438 int ret;
3439 int key_type = BTRFS_DIR_ITEM_KEY;
3440
3441again:
3442 min_key = 0;
3443 max_key = 0;
3444 while (1) {
3445 ret = log_dir_items(trans, root, inode, path,
3446 dst_path, key_type, ctx, min_key,
3447 &max_key);
3448 if (ret)
3449 return ret;
3450 if (max_key == (u64)-1)
3451 break;
3452 min_key = max_key + 1;
3453 }
3454
3455 if (key_type == BTRFS_DIR_ITEM_KEY) {
3456 key_type = BTRFS_DIR_INDEX_KEY;
3457 goto again;
3458 }
3459 return 0;
3460}
3461
3462/*
3463 * a helper function to drop items from the log before we relog an
3464 * inode. max_key_type indicates the highest item type to remove.
3465 * This cannot be run for file data extents because it does not
3466 * free the extents they point to.
3467 */
3468static int drop_objectid_items(struct btrfs_trans_handle *trans,
3469 struct btrfs_root *log,
3470 struct btrfs_path *path,
3471 u64 objectid, int max_key_type)
3472{
3473 int ret;
3474 struct btrfs_key key;
3475 struct btrfs_key found_key;
3476 int start_slot;
3477
3478 key.objectid = objectid;
3479 key.type = max_key_type;
3480 key.offset = (u64)-1;
3481
3482 while (1) {
3483 ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
3484 BUG_ON(ret == 0); /* Logic error */
3485 if (ret < 0)
3486 break;
3487
3488 if (path->slots[0] == 0)
3489 break;
3490
3491 path->slots[0]--;
3492 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
3493 path->slots[0]);
3494
3495 if (found_key.objectid != objectid)
3496 break;
3497
3498 found_key.offset = 0;
3499 found_key.type = 0;
3500 ret = btrfs_bin_search(path->nodes[0], &found_key, 0,
3501 &start_slot);
3502
3503 ret = btrfs_del_items(trans, log, path, start_slot,
3504 path->slots[0] - start_slot + 1);
3505 /*
3506 * If start slot isn't 0 then we don't need to re-search, we've
3507 * found the last guy with the objectid in this tree.
3508 */
3509 if (ret || start_slot != 0)
3510 break;
3511 btrfs_release_path(path);
3512 }
3513 btrfs_release_path(path);
3514 if (ret > 0)
3515 ret = 0;
3516 return ret;
3517}
3518
3519static void fill_inode_item(struct btrfs_trans_handle *trans,
3520 struct extent_buffer *leaf,
3521 struct btrfs_inode_item *item,
3522 struct inode *inode, int log_inode_only,
3523 u64 logged_isize)
3524{
3525 struct btrfs_map_token token;
3526
3527 btrfs_init_map_token(&token);
3528
3529 if (log_inode_only) {
3530 /* set the generation to zero so the recover code
3531 * can tell the difference between an logging
3532 * just to say 'this inode exists' and a logging
3533 * to say 'update this inode with these values'
3534 */
3535 btrfs_set_token_inode_generation(leaf, item, 0, &token);
3536 btrfs_set_token_inode_size(leaf, item, logged_isize, &token);
3537 } else {
3538 btrfs_set_token_inode_generation(leaf, item,
3539 BTRFS_I(inode)->generation,
3540 &token);
3541 btrfs_set_token_inode_size(leaf, item, inode->i_size, &token);
3542 }
3543
3544 btrfs_set_token_inode_uid(leaf, item, i_uid_read(inode), &token);
3545 btrfs_set_token_inode_gid(leaf, item, i_gid_read(inode), &token);
3546 btrfs_set_token_inode_mode(leaf, item, inode->i_mode, &token);
3547 btrfs_set_token_inode_nlink(leaf, item, inode->i_nlink, &token);
3548
3549 btrfs_set_token_timespec_sec(leaf, &item->atime,
3550 inode->i_atime.tv_sec, &token);
3551 btrfs_set_token_timespec_nsec(leaf, &item->atime,
3552 inode->i_atime.tv_nsec, &token);
3553
3554 btrfs_set_token_timespec_sec(leaf, &item->mtime,
3555 inode->i_mtime.tv_sec, &token);
3556 btrfs_set_token_timespec_nsec(leaf, &item->mtime,
3557 inode->i_mtime.tv_nsec, &token);
3558
3559 btrfs_set_token_timespec_sec(leaf, &item->ctime,
3560 inode->i_ctime.tv_sec, &token);
3561 btrfs_set_token_timespec_nsec(leaf, &item->ctime,
3562 inode->i_ctime.tv_nsec, &token);
3563
3564 btrfs_set_token_inode_nbytes(leaf, item, inode_get_bytes(inode),
3565 &token);
3566
3567 btrfs_set_token_inode_sequence(leaf, item, inode->i_version, &token);
3568 btrfs_set_token_inode_transid(leaf, item, trans->transid, &token);
3569 btrfs_set_token_inode_rdev(leaf, item, inode->i_rdev, &token);
3570 btrfs_set_token_inode_flags(leaf, item, BTRFS_I(inode)->flags, &token);
3571 btrfs_set_token_inode_block_group(leaf, item, 0, &token);
3572}
3573
3574static int log_inode_item(struct btrfs_trans_handle *trans,
3575 struct btrfs_root *log, struct btrfs_path *path,
3576 struct inode *inode)
3577{
3578 struct btrfs_inode_item *inode_item;
3579 int ret;
3580
3581 ret = btrfs_insert_empty_item(trans, log, path,
3582 &BTRFS_I(inode)->location,
3583 sizeof(*inode_item));
3584 if (ret && ret != -EEXIST)
3585 return ret;
3586 inode_item = btrfs_item_ptr(path->nodes[0], path->slots[0],
3587 struct btrfs_inode_item);
3588 fill_inode_item(trans, path->nodes[0], inode_item, inode, 0, 0);
3589 btrfs_release_path(path);
3590 return 0;
3591}
3592
3593static noinline int copy_items(struct btrfs_trans_handle *trans,
3594 struct inode *inode,
3595 struct btrfs_path *dst_path,
3596 struct btrfs_path *src_path, u64 *last_extent,
3597 int start_slot, int nr, int inode_only,
3598 u64 logged_isize)
3599{
3600 unsigned long src_offset;
3601 unsigned long dst_offset;
3602 struct btrfs_root *log = BTRFS_I(inode)->root->log_root;
3603 struct btrfs_file_extent_item *extent;
3604 struct btrfs_inode_item *inode_item;
3605 struct extent_buffer *src = src_path->nodes[0];
3606 struct btrfs_key first_key, last_key, key;
3607 int ret;
3608 struct btrfs_key *ins_keys;
3609 u32 *ins_sizes;
3610 char *ins_data;
3611 int i;
3612 struct list_head ordered_sums;
3613 int skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
3614 bool has_extents = false;
3615 bool need_find_last_extent = true;
3616 bool done = false;
3617
3618 INIT_LIST_HEAD(&ordered_sums);
3619
3620 ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
3621 nr * sizeof(u32), GFP_NOFS);
3622 if (!ins_data)
3623 return -ENOMEM;
3624
3625 first_key.objectid = (u64)-1;
3626
3627 ins_sizes = (u32 *)ins_data;
3628 ins_keys = (struct btrfs_key *)(ins_data + nr * sizeof(u32));
3629
3630 for (i = 0; i < nr; i++) {
3631 ins_sizes[i] = btrfs_item_size_nr(src, i + start_slot);
3632 btrfs_item_key_to_cpu(src, ins_keys + i, i + start_slot);
3633 }
3634 ret = btrfs_insert_empty_items(trans, log, dst_path,
3635 ins_keys, ins_sizes, nr);
3636 if (ret) {
3637 kfree(ins_data);
3638 return ret;
3639 }
3640
3641 for (i = 0; i < nr; i++, dst_path->slots[0]++) {
3642 dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
3643 dst_path->slots[0]);
3644
3645 src_offset = btrfs_item_ptr_offset(src, start_slot + i);
3646
3647 if ((i == (nr - 1)))
3648 last_key = ins_keys[i];
3649
3650 if (ins_keys[i].type == BTRFS_INODE_ITEM_KEY) {
3651 inode_item = btrfs_item_ptr(dst_path->nodes[0],
3652 dst_path->slots[0],
3653 struct btrfs_inode_item);
3654 fill_inode_item(trans, dst_path->nodes[0], inode_item,
3655 inode, inode_only == LOG_INODE_EXISTS,
3656 logged_isize);
3657 } else {
3658 copy_extent_buffer(dst_path->nodes[0], src, dst_offset,
3659 src_offset, ins_sizes[i]);
3660 }
3661
3662 /*
3663 * We set need_find_last_extent here in case we know we were
3664 * processing other items and then walk into the first extent in
3665 * the inode. If we don't hit an extent then nothing changes,
3666 * we'll do the last search the next time around.
3667 */
3668 if (ins_keys[i].type == BTRFS_EXTENT_DATA_KEY) {
3669 has_extents = true;
3670 if (first_key.objectid == (u64)-1)
3671 first_key = ins_keys[i];
3672 } else {
3673 need_find_last_extent = false;
3674 }
3675
3676 /* take a reference on file data extents so that truncates
3677 * or deletes of this inode don't have to relog the inode
3678 * again
3679 */
3680 if (ins_keys[i].type == BTRFS_EXTENT_DATA_KEY &&
3681 !skip_csum) {
3682 int found_type;
3683 extent = btrfs_item_ptr(src, start_slot + i,
3684 struct btrfs_file_extent_item);
3685
3686 if (btrfs_file_extent_generation(src, extent) < trans->transid)
3687 continue;
3688
3689 found_type = btrfs_file_extent_type(src, extent);
3690 if (found_type == BTRFS_FILE_EXTENT_REG) {
3691 u64 ds, dl, cs, cl;
3692 ds = btrfs_file_extent_disk_bytenr(src,
3693 extent);
3694 /* ds == 0 is a hole */
3695 if (ds == 0)
3696 continue;
3697
3698 dl = btrfs_file_extent_disk_num_bytes(src,
3699 extent);
3700 cs = btrfs_file_extent_offset(src, extent);
3701 cl = btrfs_file_extent_num_bytes(src,
3702 extent);
3703 if (btrfs_file_extent_compression(src,
3704 extent)) {
3705 cs = 0;
3706 cl = dl;
3707 }
3708
3709 ret = btrfs_lookup_csums_range(
3710 log->fs_info->csum_root,
3711 ds + cs, ds + cs + cl - 1,
3712 &ordered_sums, 0);
3713 if (ret) {
3714 btrfs_release_path(dst_path);
3715 kfree(ins_data);
3716 return ret;
3717 }
3718 }
3719 }
3720 }
3721
3722 btrfs_mark_buffer_dirty(dst_path->nodes[0]);
3723 btrfs_release_path(dst_path);
3724 kfree(ins_data);
3725
3726 /*
3727 * we have to do this after the loop above to avoid changing the
3728 * log tree while trying to change the log tree.
3729 */
3730 ret = 0;
3731 while (!list_empty(&ordered_sums)) {
3732 struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
3733 struct btrfs_ordered_sum,
3734 list);
3735 if (!ret)
3736 ret = btrfs_csum_file_blocks(trans, log, sums);
3737 list_del(&sums->list);
3738 kfree(sums);
3739 }
3740
3741 if (!has_extents)
3742 return ret;
3743
3744 if (need_find_last_extent && *last_extent == first_key.offset) {
3745 /*
3746 * We don't have any leafs between our current one and the one
3747 * we processed before that can have file extent items for our
3748 * inode (and have a generation number smaller than our current
3749 * transaction id).
3750 */
3751 need_find_last_extent = false;
3752 }
3753
3754 /*
3755 * Because we use btrfs_search_forward we could skip leaves that were
3756 * not modified and then assume *last_extent is valid when it really
3757 * isn't. So back up to the previous leaf and read the end of the last
3758 * extent before we go and fill in holes.
3759 */
3760 if (need_find_last_extent) {
3761 u64 len;
3762
3763 ret = btrfs_prev_leaf(BTRFS_I(inode)->root, src_path);
3764 if (ret < 0)
3765 return ret;
3766 if (ret)
3767 goto fill_holes;
3768 if (src_path->slots[0])
3769 src_path->slots[0]--;
3770 src = src_path->nodes[0];
3771 btrfs_item_key_to_cpu(src, &key, src_path->slots[0]);
3772 if (key.objectid != btrfs_ino(inode) ||
3773 key.type != BTRFS_EXTENT_DATA_KEY)
3774 goto fill_holes;
3775 extent = btrfs_item_ptr(src, src_path->slots[0],
3776 struct btrfs_file_extent_item);
3777 if (btrfs_file_extent_type(src, extent) ==
3778 BTRFS_FILE_EXTENT_INLINE) {
3779 len = btrfs_file_extent_inline_len(src,
3780 src_path->slots[0],
3781 extent);
3782 *last_extent = ALIGN(key.offset + len,
3783 log->sectorsize);
3784 } else {
3785 len = btrfs_file_extent_num_bytes(src, extent);
3786 *last_extent = key.offset + len;
3787 }
3788 }
3789fill_holes:
3790 /* So we did prev_leaf, now we need to move to the next leaf, but a few
3791 * things could have happened
3792 *
3793 * 1) A merge could have happened, so we could currently be on a leaf
3794 * that holds what we were copying in the first place.
3795 * 2) A split could have happened, and now not all of the items we want
3796 * are on the same leaf.
3797 *
3798 * So we need to adjust how we search for holes, we need to drop the
3799 * path and re-search for the first extent key we found, and then walk
3800 * forward until we hit the last one we copied.
3801 */
3802 if (need_find_last_extent) {
3803 /* btrfs_prev_leaf could return 1 without releasing the path */
3804 btrfs_release_path(src_path);
3805 ret = btrfs_search_slot(NULL, BTRFS_I(inode)->root, &first_key,
3806 src_path, 0, 0);
3807 if (ret < 0)
3808 return ret;
3809 ASSERT(ret == 0);
3810 src = src_path->nodes[0];
3811 i = src_path->slots[0];
3812 } else {
3813 i = start_slot;
3814 }
3815
3816 /*
3817 * Ok so here we need to go through and fill in any holes we may have
3818 * to make sure that holes are punched for those areas in case they had
3819 * extents previously.
3820 */
3821 while (!done) {
3822 u64 offset, len;
3823 u64 extent_end;
3824
3825 if (i >= btrfs_header_nritems(src_path->nodes[0])) {
3826 ret = btrfs_next_leaf(BTRFS_I(inode)->root, src_path);
3827 if (ret < 0)
3828 return ret;
3829 ASSERT(ret == 0);
3830 src = src_path->nodes[0];
3831 i = 0;
3832 }
3833
3834 btrfs_item_key_to_cpu(src, &key, i);
3835 if (!btrfs_comp_cpu_keys(&key, &last_key))
3836 done = true;
3837 if (key.objectid != btrfs_ino(inode) ||
3838 key.type != BTRFS_EXTENT_DATA_KEY) {
3839 i++;
3840 continue;
3841 }
3842 extent = btrfs_item_ptr(src, i, struct btrfs_file_extent_item);
3843 if (btrfs_file_extent_type(src, extent) ==
3844 BTRFS_FILE_EXTENT_INLINE) {
3845 len = btrfs_file_extent_inline_len(src, i, extent);
3846 extent_end = ALIGN(key.offset + len, log->sectorsize);
3847 } else {
3848 len = btrfs_file_extent_num_bytes(src, extent);
3849 extent_end = key.offset + len;
3850 }
3851 i++;
3852
3853 if (*last_extent == key.offset) {
3854 *last_extent = extent_end;
3855 continue;
3856 }
3857 offset = *last_extent;
3858 len = key.offset - *last_extent;
3859 ret = btrfs_insert_file_extent(trans, log, btrfs_ino(inode),
3860 offset, 0, 0, len, 0, len, 0,
3861 0, 0);
3862 if (ret)
3863 break;
3864 *last_extent = extent_end;
3865 }
3866 /*
3867 * Need to let the callers know we dropped the path so they should
3868 * re-search.
3869 */
3870 if (!ret && need_find_last_extent)
3871 ret = 1;
3872 return ret;
3873}
3874
3875static int extent_cmp(void *priv, struct list_head *a, struct list_head *b)
3876{
3877 struct extent_map *em1, *em2;
3878
3879 em1 = list_entry(a, struct extent_map, list);
3880 em2 = list_entry(b, struct extent_map, list);
3881
3882 if (em1->start < em2->start)
3883 return -1;
3884 else if (em1->start > em2->start)
3885 return 1;
3886 return 0;
3887}
3888
3889static int wait_ordered_extents(struct btrfs_trans_handle *trans,
3890 struct inode *inode,
3891 struct btrfs_root *root,
3892 const struct extent_map *em,
3893 const struct list_head *logged_list,
3894 bool *ordered_io_error)
3895{
3896 struct btrfs_ordered_extent *ordered;
3897 struct btrfs_root *log = root->log_root;
3898 u64 mod_start = em->mod_start;
3899 u64 mod_len = em->mod_len;
3900 const bool skip_csum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
3901 u64 csum_offset;
3902 u64 csum_len;
3903 LIST_HEAD(ordered_sums);
3904 int ret = 0;
3905
3906 *ordered_io_error = false;
3907
3908 if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags) ||
3909 em->block_start == EXTENT_MAP_HOLE)
3910 return 0;
3911
3912 /*
3913 * Wait far any ordered extent that covers our extent map. If it
3914 * finishes without an error, first check and see if our csums are on
3915 * our outstanding ordered extents.
3916 */
3917 list_for_each_entry(ordered, logged_list, log_list) {
3918 struct btrfs_ordered_sum *sum;
3919
3920 if (!mod_len)
3921 break;
3922
3923 if (ordered->file_offset + ordered->len <= mod_start ||
3924 mod_start + mod_len <= ordered->file_offset)
3925 continue;
3926
3927 if (!test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags) &&
3928 !test_bit(BTRFS_ORDERED_IOERR, &ordered->flags) &&
3929 !test_bit(BTRFS_ORDERED_DIRECT, &ordered->flags)) {
3930 const u64 start = ordered->file_offset;
3931 const u64 end = ordered->file_offset + ordered->len - 1;
3932
3933 WARN_ON(ordered->inode != inode);
3934 filemap_fdatawrite_range(inode->i_mapping, start, end);
3935 }
3936
3937 wait_event(ordered->wait,
3938 (test_bit(BTRFS_ORDERED_IO_DONE, &ordered->flags) ||
3939 test_bit(BTRFS_ORDERED_IOERR, &ordered->flags)));
3940
3941 if (test_bit(BTRFS_ORDERED_IOERR, &ordered->flags)) {
3942 /*
3943 * Clear the AS_EIO/AS_ENOSPC flags from the inode's
3944 * i_mapping flags, so that the next fsync won't get
3945 * an outdated io error too.
3946 */
3947 btrfs_inode_check_errors(inode);
3948 *ordered_io_error = true;
3949 break;
3950 }
3951 /*
3952 * We are going to copy all the csums on this ordered extent, so
3953 * go ahead and adjust mod_start and mod_len in case this
3954 * ordered extent has already been logged.
3955 */
3956 if (ordered->file_offset > mod_start) {
3957 if (ordered->file_offset + ordered->len >=
3958 mod_start + mod_len)
3959 mod_len = ordered->file_offset - mod_start;
3960 /*
3961 * If we have this case
3962 *
3963 * |--------- logged extent ---------|
3964 * |----- ordered extent ----|
3965 *
3966 * Just don't mess with mod_start and mod_len, we'll
3967 * just end up logging more csums than we need and it
3968 * will be ok.
3969 */
3970 } else {
3971 if (ordered->file_offset + ordered->len <
3972 mod_start + mod_len) {
3973 mod_len = (mod_start + mod_len) -
3974 (ordered->file_offset + ordered->len);
3975 mod_start = ordered->file_offset +
3976 ordered->len;
3977 } else {
3978 mod_len = 0;
3979 }
3980 }
3981
3982 if (skip_csum)
3983 continue;
3984
3985 /*
3986 * To keep us from looping for the above case of an ordered
3987 * extent that falls inside of the logged extent.
3988 */
3989 if (test_and_set_bit(BTRFS_ORDERED_LOGGED_CSUM,
3990 &ordered->flags))
3991 continue;
3992
3993 list_for_each_entry(sum, &ordered->list, list) {
3994 ret = btrfs_csum_file_blocks(trans, log, sum);
3995 if (ret)
3996 break;
3997 }
3998 }
3999
4000 if (*ordered_io_error || !mod_len || ret || skip_csum)
4001 return ret;
4002
4003 if (em->compress_type) {
4004 csum_offset = 0;
4005 csum_len = max(em->block_len, em->orig_block_len);
4006 } else {
4007 csum_offset = mod_start - em->start;
4008 csum_len = mod_len;
4009 }
4010
4011 /* block start is already adjusted for the file extent offset. */
4012 ret = btrfs_lookup_csums_range(log->fs_info->csum_root,
4013 em->block_start + csum_offset,
4014 em->block_start + csum_offset +
4015 csum_len - 1, &ordered_sums, 0);
4016 if (ret)
4017 return ret;
4018
4019 while (!list_empty(&ordered_sums)) {
4020 struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
4021 struct btrfs_ordered_sum,
4022 list);
4023 if (!ret)
4024 ret = btrfs_csum_file_blocks(trans, log, sums);
4025 list_del(&sums->list);
4026 kfree(sums);
4027 }
4028
4029 return ret;
4030}
4031
4032static int log_one_extent(struct btrfs_trans_handle *trans,
4033 struct inode *inode, struct btrfs_root *root,
4034 const struct extent_map *em,
4035 struct btrfs_path *path,
4036 const struct list_head *logged_list,
4037 struct btrfs_log_ctx *ctx)
4038{
4039 struct btrfs_root *log = root->log_root;
4040 struct btrfs_file_extent_item *fi;
4041 struct extent_buffer *leaf;
4042 struct btrfs_map_token token;
4043 struct btrfs_key key;
4044 u64 extent_offset = em->start - em->orig_start;
4045 u64 block_len;
4046 int ret;
4047 int extent_inserted = 0;
4048 bool ordered_io_err = false;
4049
4050 ret = wait_ordered_extents(trans, inode, root, em, logged_list,
4051 &ordered_io_err);
4052 if (ret)
4053 return ret;
4054
4055 if (ordered_io_err) {
4056 ctx->io_err = -EIO;
4057 return 0;
4058 }
4059
4060 btrfs_init_map_token(&token);
4061
4062 ret = __btrfs_drop_extents(trans, log, inode, path, em->start,
4063 em->start + em->len, NULL, 0, 1,
4064 sizeof(*fi), &extent_inserted);
4065 if (ret)
4066 return ret;
4067
4068 if (!extent_inserted) {
4069 key.objectid = btrfs_ino(inode);
4070 key.type = BTRFS_EXTENT_DATA_KEY;
4071 key.offset = em->start;
4072
4073 ret = btrfs_insert_empty_item(trans, log, path, &key,
4074 sizeof(*fi));
4075 if (ret)
4076 return ret;
4077 }
4078 leaf = path->nodes[0];
4079 fi = btrfs_item_ptr(leaf, path->slots[0],
4080 struct btrfs_file_extent_item);
4081
4082 btrfs_set_token_file_extent_generation(leaf, fi, trans->transid,
4083 &token);
4084 if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
4085 btrfs_set_token_file_extent_type(leaf, fi,
4086 BTRFS_FILE_EXTENT_PREALLOC,
4087 &token);
4088 else
4089 btrfs_set_token_file_extent_type(leaf, fi,
4090 BTRFS_FILE_EXTENT_REG,
4091 &token);
4092
4093 block_len = max(em->block_len, em->orig_block_len);
4094 if (em->compress_type != BTRFS_COMPRESS_NONE) {
4095 btrfs_set_token_file_extent_disk_bytenr(leaf, fi,
4096 em->block_start,
4097 &token);
4098 btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, block_len,
4099 &token);
4100 } else if (em->block_start < EXTENT_MAP_LAST_BYTE) {
4101 btrfs_set_token_file_extent_disk_bytenr(leaf, fi,
4102 em->block_start -
4103 extent_offset, &token);
4104 btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, block_len,
4105 &token);
4106 } else {
4107 btrfs_set_token_file_extent_disk_bytenr(leaf, fi, 0, &token);
4108 btrfs_set_token_file_extent_disk_num_bytes(leaf, fi, 0,
4109 &token);
4110 }
4111
4112 btrfs_set_token_file_extent_offset(leaf, fi, extent_offset, &token);
4113 btrfs_set_token_file_extent_num_bytes(leaf, fi, em->len, &token);
4114 btrfs_set_token_file_extent_ram_bytes(leaf, fi, em->ram_bytes, &token);
4115 btrfs_set_token_file_extent_compression(leaf, fi, em->compress_type,
4116 &token);
4117 btrfs_set_token_file_extent_encryption(leaf, fi, 0, &token);
4118 btrfs_set_token_file_extent_other_encoding(leaf, fi, 0, &token);
4119 btrfs_mark_buffer_dirty(leaf);
4120
4121 btrfs_release_path(path);
4122
4123 return ret;
4124}
4125
4126static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
4127 struct btrfs_root *root,
4128 struct inode *inode,
4129 struct btrfs_path *path,
4130 struct list_head *logged_list,
4131 struct btrfs_log_ctx *ctx,
4132 const u64 start,
4133 const u64 end)
4134{
4135 struct extent_map *em, *n;
4136 struct list_head extents;
4137 struct extent_map_tree *tree = &BTRFS_I(inode)->extent_tree;
4138 u64 test_gen;
4139 int ret = 0;
4140 int num = 0;
4141
4142 INIT_LIST_HEAD(&extents);
4143
4144 write_lock(&tree->lock);
4145 test_gen = root->fs_info->last_trans_committed;
4146
4147 list_for_each_entry_safe(em, n, &tree->modified_extents, list) {
4148 list_del_init(&em->list);
4149
4150 /*
4151 * Just an arbitrary number, this can be really CPU intensive
4152 * once we start getting a lot of extents, and really once we
4153 * have a bunch of extents we just want to commit since it will
4154 * be faster.
4155 */
4156 if (++num > 32768) {
4157 list_del_init(&tree->modified_extents);
4158 ret = -EFBIG;
4159 goto process;
4160 }
4161
4162 if (em->generation <= test_gen)
4163 continue;
4164 /* Need a ref to keep it from getting evicted from cache */
4165 atomic_inc(&em->refs);
4166 set_bit(EXTENT_FLAG_LOGGING, &em->flags);
4167 list_add_tail(&em->list, &extents);
4168 num++;
4169 }
4170
4171 list_sort(NULL, &extents, extent_cmp);
4172 /*
4173 * Collect any new ordered extents within the range. This is to
4174 * prevent logging file extent items without waiting for the disk
4175 * location they point to being written. We do this only to deal
4176 * with races against concurrent lockless direct IO writes.
4177 */
4178 btrfs_get_logged_extents(inode, logged_list, start, end);
4179process:
4180 while (!list_empty(&extents)) {
4181 em = list_entry(extents.next, struct extent_map, list);
4182
4183 list_del_init(&em->list);
4184
4185 /*
4186 * If we had an error we just need to delete everybody from our
4187 * private list.
4188 */
4189 if (ret) {
4190 clear_em_logging(tree, em);
4191 free_extent_map(em);
4192 continue;
4193 }
4194
4195 write_unlock(&tree->lock);
4196
4197 ret = log_one_extent(trans, inode, root, em, path, logged_list,
4198 ctx);
4199 write_lock(&tree->lock);
4200 clear_em_logging(tree, em);
4201 free_extent_map(em);
4202 }
4203 WARN_ON(!list_empty(&extents));
4204 write_unlock(&tree->lock);
4205
4206 btrfs_release_path(path);
4207 return ret;
4208}
4209
4210static int logged_inode_size(struct btrfs_root *log, struct inode *inode,
4211 struct btrfs_path *path, u64 *size_ret)
4212{
4213 struct btrfs_key key;
4214 int ret;
4215
4216 key.objectid = btrfs_ino(inode);
4217 key.type = BTRFS_INODE_ITEM_KEY;
4218 key.offset = 0;
4219
4220 ret = btrfs_search_slot(NULL, log, &key, path, 0, 0);
4221 if (ret < 0) {
4222 return ret;
4223 } else if (ret > 0) {
4224 *size_ret = 0;
4225 } else {
4226 struct btrfs_inode_item *item;
4227
4228 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
4229 struct btrfs_inode_item);
4230 *size_ret = btrfs_inode_size(path->nodes[0], item);
4231 }
4232
4233 btrfs_release_path(path);
4234 return 0;
4235}
4236
4237/*
4238 * At the moment we always log all xattrs. This is to figure out at log replay
4239 * time which xattrs must have their deletion replayed. If a xattr is missing
4240 * in the log tree and exists in the fs/subvol tree, we delete it. This is
4241 * because if a xattr is deleted, the inode is fsynced and a power failure
4242 * happens, causing the log to be replayed the next time the fs is mounted,
4243 * we want the xattr to not exist anymore (same behaviour as other filesystems
4244 * with a journal, ext3/4, xfs, f2fs, etc).
4245 */
4246static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
4247 struct btrfs_root *root,
4248 struct inode *inode,
4249 struct btrfs_path *path,
4250 struct btrfs_path *dst_path)
4251{
4252 int ret;
4253 struct btrfs_key key;
4254 const u64 ino = btrfs_ino(inode);
4255 int ins_nr = 0;
4256 int start_slot = 0;
4257
4258 key.objectid = ino;
4259 key.type = BTRFS_XATTR_ITEM_KEY;
4260 key.offset = 0;
4261
4262 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4263 if (ret < 0)
4264 return ret;
4265
4266 while (true) {
4267 int slot = path->slots[0];
4268 struct extent_buffer *leaf = path->nodes[0];
4269 int nritems = btrfs_header_nritems(leaf);
4270
4271 if (slot >= nritems) {
4272 if (ins_nr > 0) {
4273 u64 last_extent = 0;
4274
4275 ret = copy_items(trans, inode, dst_path, path,
4276 &last_extent, start_slot,
4277 ins_nr, 1, 0);
4278 /* can't be 1, extent items aren't processed */
4279 ASSERT(ret <= 0);
4280 if (ret < 0)
4281 return ret;
4282 ins_nr = 0;
4283 }
4284 ret = btrfs_next_leaf(root, path);
4285 if (ret < 0)
4286 return ret;
4287 else if (ret > 0)
4288 break;
4289 continue;
4290 }
4291
4292 btrfs_item_key_to_cpu(leaf, &key, slot);
4293 if (key.objectid != ino || key.type != BTRFS_XATTR_ITEM_KEY)
4294 break;
4295
4296 if (ins_nr == 0)
4297 start_slot = slot;
4298 ins_nr++;
4299 path->slots[0]++;
4300 cond_resched();
4301 }
4302 if (ins_nr > 0) {
4303 u64 last_extent = 0;
4304
4305 ret = copy_items(trans, inode, dst_path, path,
4306 &last_extent, start_slot,
4307 ins_nr, 1, 0);
4308 /* can't be 1, extent items aren't processed */
4309 ASSERT(ret <= 0);
4310 if (ret < 0)
4311 return ret;
4312 }
4313
4314 return 0;
4315}
4316
4317/*
4318 * If the no holes feature is enabled we need to make sure any hole between the
4319 * last extent and the i_size of our inode is explicitly marked in the log. This
4320 * is to make sure that doing something like:
4321 *
4322 * 1) create file with 128Kb of data
4323 * 2) truncate file to 64Kb
4324 * 3) truncate file to 256Kb
4325 * 4) fsync file
4326 * 5) <crash/power failure>
4327 * 6) mount fs and trigger log replay
4328 *
4329 * Will give us a file with a size of 256Kb, the first 64Kb of data match what
4330 * the file had in its first 64Kb of data at step 1 and the last 192Kb of the
4331 * file correspond to a hole. The presence of explicit holes in a log tree is
4332 * what guarantees that log replay will remove/adjust file extent items in the
4333 * fs/subvol tree.
4334 *
4335 * Here we do not need to care about holes between extents, that is already done
4336 * by copy_items(). We also only need to do this in the full sync path, where we
4337 * lookup for extents from the fs/subvol tree only. In the fast path case, we
4338 * lookup the list of modified extent maps and if any represents a hole, we
4339 * insert a corresponding extent representing a hole in the log tree.
4340 */
4341static int btrfs_log_trailing_hole(struct btrfs_trans_handle *trans,
4342 struct btrfs_root *root,
4343 struct inode *inode,
4344 struct btrfs_path *path)
4345{
4346 int ret;
4347 struct btrfs_key key;
4348 u64 hole_start;
4349 u64 hole_size;
4350 struct extent_buffer *leaf;
4351 struct btrfs_root *log = root->log_root;
4352 const u64 ino = btrfs_ino(inode);
4353 const u64 i_size = i_size_read(inode);
4354
4355 if (!btrfs_fs_incompat(root->fs_info, NO_HOLES))
4356 return 0;
4357
4358 key.objectid = ino;
4359 key.type = BTRFS_EXTENT_DATA_KEY;
4360 key.offset = (u64)-1;
4361
4362 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4363 ASSERT(ret != 0);
4364 if (ret < 0)
4365 return ret;
4366
4367 ASSERT(path->slots[0] > 0);
4368 path->slots[0]--;
4369 leaf = path->nodes[0];
4370 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
4371
4372 if (key.objectid != ino || key.type != BTRFS_EXTENT_DATA_KEY) {
4373 /* inode does not have any extents */
4374 hole_start = 0;
4375 hole_size = i_size;
4376 } else {
4377 struct btrfs_file_extent_item *extent;
4378 u64 len;
4379
4380 /*
4381 * If there's an extent beyond i_size, an explicit hole was
4382 * already inserted by copy_items().
4383 */
4384 if (key.offset >= i_size)
4385 return 0;
4386
4387 extent = btrfs_item_ptr(leaf, path->slots[0],
4388 struct btrfs_file_extent_item);
4389
4390 if (btrfs_file_extent_type(leaf, extent) ==
4391 BTRFS_FILE_EXTENT_INLINE) {
4392 len = btrfs_file_extent_inline_len(leaf,
4393 path->slots[0],
4394 extent);
4395 ASSERT(len == i_size);
4396 return 0;
4397 }
4398
4399 len = btrfs_file_extent_num_bytes(leaf, extent);
4400 /* Last extent goes beyond i_size, no need to log a hole. */
4401 if (key.offset + len > i_size)
4402 return 0;
4403 hole_start = key.offset + len;
4404 hole_size = i_size - hole_start;
4405 }
4406 btrfs_release_path(path);
4407
4408 /* Last extent ends at i_size. */
4409 if (hole_size == 0)
4410 return 0;
4411
4412 hole_size = ALIGN(hole_size, root->sectorsize);
4413 ret = btrfs_insert_file_extent(trans, log, ino, hole_start, 0, 0,
4414 hole_size, 0, hole_size, 0, 0, 0);
4415 return ret;
4416}
4417
4418/*
4419 * When we are logging a new inode X, check if it doesn't have a reference that
4420 * matches the reference from some other inode Y created in a past transaction
4421 * and that was renamed in the current transaction. If we don't do this, then at
4422 * log replay time we can lose inode Y (and all its files if it's a directory):
4423 *
4424 * mkdir /mnt/x
4425 * echo "hello world" > /mnt/x/foobar
4426 * sync
4427 * mv /mnt/x /mnt/y
4428 * mkdir /mnt/x # or touch /mnt/x
4429 * xfs_io -c fsync /mnt/x
4430 * <power fail>
4431 * mount fs, trigger log replay
4432 *
4433 * After the log replay procedure, we would lose the first directory and all its
4434 * files (file foobar).
4435 * For the case where inode Y is not a directory we simply end up losing it:
4436 *
4437 * echo "123" > /mnt/foo
4438 * sync
4439 * mv /mnt/foo /mnt/bar
4440 * echo "abc" > /mnt/foo
4441 * xfs_io -c fsync /mnt/foo
4442 * <power fail>
4443 *
4444 * We also need this for cases where a snapshot entry is replaced by some other
4445 * entry (file or directory) otherwise we end up with an unreplayable log due to
4446 * attempts to delete the snapshot entry (entry of type BTRFS_ROOT_ITEM_KEY) as
4447 * if it were a regular entry:
4448 *
4449 * mkdir /mnt/x
4450 * btrfs subvolume snapshot /mnt /mnt/x/snap
4451 * btrfs subvolume delete /mnt/x/snap
4452 * rmdir /mnt/x
4453 * mkdir /mnt/x
4454 * fsync /mnt/x or fsync some new file inside it
4455 * <power fail>
4456 *
4457 * The snapshot delete, rmdir of x, mkdir of a new x and the fsync all happen in
4458 * the same transaction.
4459 */
4460static int btrfs_check_ref_name_override(struct extent_buffer *eb,
4461 const int slot,
4462 const struct btrfs_key *key,
4463 struct inode *inode)
4464{
4465 int ret;
4466 struct btrfs_path *search_path;
4467 char *name = NULL;
4468 u32 name_len = 0;
4469 u32 item_size = btrfs_item_size_nr(eb, slot);
4470 u32 cur_offset = 0;
4471 unsigned long ptr = btrfs_item_ptr_offset(eb, slot);
4472
4473 search_path = btrfs_alloc_path();
4474 if (!search_path)
4475 return -ENOMEM;
4476 search_path->search_commit_root = 1;
4477 search_path->skip_locking = 1;
4478
4479 while (cur_offset < item_size) {
4480 u64 parent;
4481 u32 this_name_len;
4482 u32 this_len;
4483 unsigned long name_ptr;
4484 struct btrfs_dir_item *di;
4485
4486 if (key->type == BTRFS_INODE_REF_KEY) {
4487 struct btrfs_inode_ref *iref;
4488
4489 iref = (struct btrfs_inode_ref *)(ptr + cur_offset);
4490 parent = key->offset;
4491 this_name_len = btrfs_inode_ref_name_len(eb, iref);
4492 name_ptr = (unsigned long)(iref + 1);
4493 this_len = sizeof(*iref) + this_name_len;
4494 } else {
4495 struct btrfs_inode_extref *extref;
4496
4497 extref = (struct btrfs_inode_extref *)(ptr +
4498 cur_offset);
4499 parent = btrfs_inode_extref_parent(eb, extref);
4500 this_name_len = btrfs_inode_extref_name_len(eb, extref);
4501 name_ptr = (unsigned long)&extref->name;
4502 this_len = sizeof(*extref) + this_name_len;
4503 }
4504
4505 if (this_name_len > name_len) {
4506 char *new_name;
4507
4508 new_name = krealloc(name, this_name_len, GFP_NOFS);
4509 if (!new_name) {
4510 ret = -ENOMEM;
4511 goto out;
4512 }
4513 name_len = this_name_len;
4514 name = new_name;
4515 }
4516
4517 read_extent_buffer(eb, name, name_ptr, this_name_len);
4518 di = btrfs_lookup_dir_item(NULL, BTRFS_I(inode)->root,
4519 search_path, parent,
4520 name, this_name_len, 0);
4521 if (di && !IS_ERR(di)) {
4522 ret = 1;
4523 goto out;
4524 } else if (IS_ERR(di)) {
4525 ret = PTR_ERR(di);
4526 goto out;
4527 }
4528 btrfs_release_path(search_path);
4529
4530 cur_offset += this_len;
4531 }
4532 ret = 0;
4533out:
4534 btrfs_free_path(search_path);
4535 kfree(name);
4536 return ret;
4537}
4538
4539/* log a single inode in the tree log.
4540 * At least one parent directory for this inode must exist in the tree
4541 * or be logged already.
4542 *
4543 * Any items from this inode changed by the current transaction are copied
4544 * to the log tree. An extra reference is taken on any extents in this
4545 * file, allowing us to avoid a whole pile of corner cases around logging
4546 * blocks that have been removed from the tree.
4547 *
4548 * See LOG_INODE_ALL and related defines for a description of what inode_only
4549 * does.
4550 *
4551 * This handles both files and directories.
4552 */
4553static int btrfs_log_inode(struct btrfs_trans_handle *trans,
4554 struct btrfs_root *root, struct inode *inode,
4555 int inode_only,
4556 const loff_t start,
4557 const loff_t end,
4558 struct btrfs_log_ctx *ctx)
4559{
4560 struct btrfs_path *path;
4561 struct btrfs_path *dst_path;
4562 struct btrfs_key min_key;
4563 struct btrfs_key max_key;
4564 struct btrfs_root *log = root->log_root;
4565 struct extent_buffer *src = NULL;
4566 LIST_HEAD(logged_list);
4567 u64 last_extent = 0;
4568 int err = 0;
4569 int ret;
4570 int nritems;
4571 int ins_start_slot = 0;
4572 int ins_nr;
4573 bool fast_search = false;
4574 u64 ino = btrfs_ino(inode);
4575 struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
4576 u64 logged_isize = 0;
4577 bool need_log_inode_item = true;
4578
4579 path = btrfs_alloc_path();
4580 if (!path)
4581 return -ENOMEM;
4582 dst_path = btrfs_alloc_path();
4583 if (!dst_path) {
4584 btrfs_free_path(path);
4585 return -ENOMEM;
4586 }
4587
4588 min_key.objectid = ino;
4589 min_key.type = BTRFS_INODE_ITEM_KEY;
4590 min_key.offset = 0;
4591
4592 max_key.objectid = ino;
4593
4594
4595 /* today the code can only do partial logging of directories */
4596 if (S_ISDIR(inode->i_mode) ||
4597 (!test_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
4598 &BTRFS_I(inode)->runtime_flags) &&
4599 inode_only == LOG_INODE_EXISTS))
4600 max_key.type = BTRFS_XATTR_ITEM_KEY;
4601 else
4602 max_key.type = (u8)-1;
4603 max_key.offset = (u64)-1;
4604
4605 /*
4606 * Only run delayed items if we are a dir or a new file.
4607 * Otherwise commit the delayed inode only, which is needed in
4608 * order for the log replay code to mark inodes for link count
4609 * fixup (create temporary BTRFS_TREE_LOG_FIXUP_OBJECTID items).
4610 */
4611 if (S_ISDIR(inode->i_mode) ||
4612 BTRFS_I(inode)->generation > root->fs_info->last_trans_committed)
4613 ret = btrfs_commit_inode_delayed_items(trans, inode);
4614 else
4615 ret = btrfs_commit_inode_delayed_inode(inode);
4616
4617 if (ret) {
4618 btrfs_free_path(path);
4619 btrfs_free_path(dst_path);
4620 return ret;
4621 }
4622
4623 mutex_lock(&BTRFS_I(inode)->log_mutex);
4624
4625 /*
4626 * Collect ordered extents only if we are logging data. This is to
4627 * ensure a subsequent request to log this inode in LOG_INODE_ALL mode
4628 * will process the ordered extents if they still exists at the time,
4629 * because when we collect them we test and set for the flag
4630 * BTRFS_ORDERED_LOGGED to prevent multiple log requests to process the
4631 * same ordered extents. The consequence for the LOG_INODE_ALL log mode
4632 * not processing the ordered extents is that we end up logging the
4633 * corresponding file extent items, based on the extent maps in the
4634 * inode's extent_map_tree's modified_list, without logging the
4635 * respective checksums (since the may still be only attached to the
4636 * ordered extents and have not been inserted in the csum tree by
4637 * btrfs_finish_ordered_io() yet).
4638 */
4639 if (inode_only == LOG_INODE_ALL)
4640 btrfs_get_logged_extents(inode, &logged_list, start, end);
4641
4642 /*
4643 * a brute force approach to making sure we get the most uptodate
4644 * copies of everything.
4645 */
4646 if (S_ISDIR(inode->i_mode)) {
4647 int max_key_type = BTRFS_DIR_LOG_INDEX_KEY;
4648
4649 if (inode_only == LOG_INODE_EXISTS)
4650 max_key_type = BTRFS_XATTR_ITEM_KEY;
4651 ret = drop_objectid_items(trans, log, path, ino, max_key_type);
4652 } else {
4653 if (inode_only == LOG_INODE_EXISTS) {
4654 /*
4655 * Make sure the new inode item we write to the log has
4656 * the same isize as the current one (if it exists).
4657 * This is necessary to prevent data loss after log
4658 * replay, and also to prevent doing a wrong expanding
4659 * truncate - for e.g. create file, write 4K into offset
4660 * 0, fsync, write 4K into offset 4096, add hard link,
4661 * fsync some other file (to sync log), power fail - if
4662 * we use the inode's current i_size, after log replay
4663 * we get a 8Kb file, with the last 4Kb extent as a hole
4664 * (zeroes), as if an expanding truncate happened,
4665 * instead of getting a file of 4Kb only.
4666 */
4667 err = logged_inode_size(log, inode, path,
4668 &logged_isize);
4669 if (err)
4670 goto out_unlock;
4671 }
4672 if (test_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
4673 &BTRFS_I(inode)->runtime_flags)) {
4674 if (inode_only == LOG_INODE_EXISTS) {
4675 max_key.type = BTRFS_XATTR_ITEM_KEY;
4676 ret = drop_objectid_items(trans, log, path, ino,
4677 max_key.type);
4678 } else {
4679 clear_bit(BTRFS_INODE_NEEDS_FULL_SYNC,
4680 &BTRFS_I(inode)->runtime_flags);
4681 clear_bit(BTRFS_INODE_COPY_EVERYTHING,
4682 &BTRFS_I(inode)->runtime_flags);
4683 while(1) {
4684 ret = btrfs_truncate_inode_items(trans,
4685 log, inode, 0, 0);
4686 if (ret != -EAGAIN)
4687 break;
4688 }
4689 }
4690 } else if (test_and_clear_bit(BTRFS_INODE_COPY_EVERYTHING,
4691 &BTRFS_I(inode)->runtime_flags) ||
4692 inode_only == LOG_INODE_EXISTS) {
4693 if (inode_only == LOG_INODE_ALL)
4694 fast_search = true;
4695 max_key.type = BTRFS_XATTR_ITEM_KEY;
4696 ret = drop_objectid_items(trans, log, path, ino,
4697 max_key.type);
4698 } else {
4699 if (inode_only == LOG_INODE_ALL)
4700 fast_search = true;
4701 goto log_extents;
4702 }
4703
4704 }
4705 if (ret) {
4706 err = ret;
4707 goto out_unlock;
4708 }
4709
4710 while (1) {
4711 ins_nr = 0;
4712 ret = btrfs_search_forward(root, &min_key,
4713 path, trans->transid);
4714 if (ret != 0)
4715 break;
4716again:
4717 /* note, ins_nr might be > 0 here, cleanup outside the loop */
4718 if (min_key.objectid != ino)
4719 break;
4720 if (min_key.type > max_key.type)
4721 break;
4722
4723 if (min_key.type == BTRFS_INODE_ITEM_KEY)
4724 need_log_inode_item = false;
4725
4726 if ((min_key.type == BTRFS_INODE_REF_KEY ||
4727 min_key.type == BTRFS_INODE_EXTREF_KEY) &&
4728 BTRFS_I(inode)->generation == trans->transid) {
4729 ret = btrfs_check_ref_name_override(path->nodes[0],
4730 path->slots[0],
4731 &min_key, inode);
4732 if (ret < 0) {
4733 err = ret;
4734 goto out_unlock;
4735 } else if (ret > 0) {
4736 err = 1;
4737 btrfs_set_log_full_commit(root->fs_info, trans);
4738 goto out_unlock;
4739 }
4740 }
4741
4742 /* Skip xattrs, we log them later with btrfs_log_all_xattrs() */
4743 if (min_key.type == BTRFS_XATTR_ITEM_KEY) {
4744 if (ins_nr == 0)
4745 goto next_slot;
4746 ret = copy_items(trans, inode, dst_path, path,
4747 &last_extent, ins_start_slot,
4748 ins_nr, inode_only, logged_isize);
4749 if (ret < 0) {
4750 err = ret;
4751 goto out_unlock;
4752 }
4753 ins_nr = 0;
4754 if (ret) {
4755 btrfs_release_path(path);
4756 continue;
4757 }
4758 goto next_slot;
4759 }
4760
4761 src = path->nodes[0];
4762 if (ins_nr && ins_start_slot + ins_nr == path->slots[0]) {
4763 ins_nr++;
4764 goto next_slot;
4765 } else if (!ins_nr) {
4766 ins_start_slot = path->slots[0];
4767 ins_nr = 1;
4768 goto next_slot;
4769 }
4770
4771 ret = copy_items(trans, inode, dst_path, path, &last_extent,
4772 ins_start_slot, ins_nr, inode_only,
4773 logged_isize);
4774 if (ret < 0) {
4775 err = ret;
4776 goto out_unlock;
4777 }
4778 if (ret) {
4779 ins_nr = 0;
4780 btrfs_release_path(path);
4781 continue;
4782 }
4783 ins_nr = 1;
4784 ins_start_slot = path->slots[0];
4785next_slot:
4786
4787 nritems = btrfs_header_nritems(path->nodes[0]);
4788 path->slots[0]++;
4789 if (path->slots[0] < nritems) {
4790 btrfs_item_key_to_cpu(path->nodes[0], &min_key,
4791 path->slots[0]);
4792 goto again;
4793 }
4794 if (ins_nr) {
4795 ret = copy_items(trans, inode, dst_path, path,
4796 &last_extent, ins_start_slot,
4797 ins_nr, inode_only, logged_isize);
4798 if (ret < 0) {
4799 err = ret;
4800 goto out_unlock;
4801 }
4802 ret = 0;
4803 ins_nr = 0;
4804 }
4805 btrfs_release_path(path);
4806
4807 if (min_key.offset < (u64)-1) {
4808 min_key.offset++;
4809 } else if (min_key.type < max_key.type) {
4810 min_key.type++;
4811 min_key.offset = 0;
4812 } else {
4813 break;
4814 }
4815 }
4816 if (ins_nr) {
4817 ret = copy_items(trans, inode, dst_path, path, &last_extent,
4818 ins_start_slot, ins_nr, inode_only,
4819 logged_isize);
4820 if (ret < 0) {
4821 err = ret;
4822 goto out_unlock;
4823 }
4824 ret = 0;
4825 ins_nr = 0;
4826 }
4827
4828 btrfs_release_path(path);
4829 btrfs_release_path(dst_path);
4830 err = btrfs_log_all_xattrs(trans, root, inode, path, dst_path);
4831 if (err)
4832 goto out_unlock;
4833 if (max_key.type >= BTRFS_EXTENT_DATA_KEY && !fast_search) {
4834 btrfs_release_path(path);
4835 btrfs_release_path(dst_path);
4836 err = btrfs_log_trailing_hole(trans, root, inode, path);
4837 if (err)
4838 goto out_unlock;
4839 }
4840log_extents:
4841 btrfs_release_path(path);
4842 btrfs_release_path(dst_path);
4843 if (need_log_inode_item) {
4844 err = log_inode_item(trans, log, dst_path, inode);
4845 if (err)
4846 goto out_unlock;
4847 }
4848 if (fast_search) {
4849 /*
4850 * Some ordered extents started by fsync might have completed
4851 * before we collected the ordered extents in logged_list, which
4852 * means they're gone, not in our logged_list nor in the inode's
4853 * ordered tree. We want the application/user space to know an
4854 * error happened while attempting to persist file data so that
4855 * it can take proper action. If such error happened, we leave
4856 * without writing to the log tree and the fsync must report the
4857 * file data write error and not commit the current transaction.
4858 */
4859 err = btrfs_inode_check_errors(inode);
4860 if (err) {
4861 ctx->io_err = err;
4862 goto out_unlock;
4863 }
4864 ret = btrfs_log_changed_extents(trans, root, inode, dst_path,
4865 &logged_list, ctx, start, end);
4866 if (ret) {
4867 err = ret;
4868 goto out_unlock;
4869 }
4870 } else if (inode_only == LOG_INODE_ALL) {
4871 struct extent_map *em, *n;
4872
4873 write_lock(&em_tree->lock);
4874 /*
4875 * We can't just remove every em if we're called for a ranged
4876 * fsync - that is, one that doesn't cover the whole possible
4877 * file range (0 to LLONG_MAX). This is because we can have
4878 * em's that fall outside the range we're logging and therefore
4879 * their ordered operations haven't completed yet
4880 * (btrfs_finish_ordered_io() not invoked yet). This means we
4881 * didn't get their respective file extent item in the fs/subvol
4882 * tree yet, and need to let the next fast fsync (one which
4883 * consults the list of modified extent maps) find the em so
4884 * that it logs a matching file extent item and waits for the
4885 * respective ordered operation to complete (if it's still
4886 * running).
4887 *
4888 * Removing every em outside the range we're logging would make
4889 * the next fast fsync not log their matching file extent items,
4890 * therefore making us lose data after a log replay.
4891 */
4892 list_for_each_entry_safe(em, n, &em_tree->modified_extents,
4893 list) {
4894 const u64 mod_end = em->mod_start + em->mod_len - 1;
4895
4896 if (em->mod_start >= start && mod_end <= end)
4897 list_del_init(&em->list);
4898 }
4899 write_unlock(&em_tree->lock);
4900 }
4901
4902 if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
4903 ret = log_directory_changes(trans, root, inode, path, dst_path,
4904 ctx);
4905 if (ret) {
4906 err = ret;
4907 goto out_unlock;
4908 }
4909 }
4910
4911 spin_lock(&BTRFS_I(inode)->lock);
4912 BTRFS_I(inode)->logged_trans = trans->transid;
4913 BTRFS_I(inode)->last_log_commit = BTRFS_I(inode)->last_sub_trans;
4914 spin_unlock(&BTRFS_I(inode)->lock);
4915out_unlock:
4916 if (unlikely(err))
4917 btrfs_put_logged_extents(&logged_list);
4918 else
4919 btrfs_submit_logged_extents(&logged_list, log);
4920 mutex_unlock(&BTRFS_I(inode)->log_mutex);
4921
4922 btrfs_free_path(path);
4923 btrfs_free_path(dst_path);
4924 return err;
4925}
4926
4927/*
4928 * Check if we must fallback to a transaction commit when logging an inode.
4929 * This must be called after logging the inode and is used only in the context
4930 * when fsyncing an inode requires the need to log some other inode - in which
4931 * case we can't lock the i_mutex of each other inode we need to log as that
4932 * can lead to deadlocks with concurrent fsync against other inodes (as we can
4933 * log inodes up or down in the hierarchy) or rename operations for example. So
4934 * we take the log_mutex of the inode after we have logged it and then check for
4935 * its last_unlink_trans value - this is safe because any task setting
4936 * last_unlink_trans must take the log_mutex and it must do this before it does
4937 * the actual unlink operation, so if we do this check before a concurrent task
4938 * sets last_unlink_trans it means we've logged a consistent version/state of
4939 * all the inode items, otherwise we are not sure and must do a transaction
4940 * commit (the concurrent task migth have only updated last_unlink_trans before
4941 * we logged the inode or it might have also done the unlink).
4942 */
4943static bool btrfs_must_commit_transaction(struct btrfs_trans_handle *trans,
4944 struct inode *inode)
4945{
4946 struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
4947 bool ret = false;
4948
4949 mutex_lock(&BTRFS_I(inode)->log_mutex);
4950 if (BTRFS_I(inode)->last_unlink_trans > fs_info->last_trans_committed) {
4951 /*
4952 * Make sure any commits to the log are forced to be full
4953 * commits.
4954 */
4955 btrfs_set_log_full_commit(fs_info, trans);
4956 ret = true;
4957 }
4958 mutex_unlock(&BTRFS_I(inode)->log_mutex);
4959
4960 return ret;
4961}
4962
4963/*
4964 * follow the dentry parent pointers up the chain and see if any
4965 * of the directories in it require a full commit before they can
4966 * be logged. Returns zero if nothing special needs to be done or 1 if
4967 * a full commit is required.
4968 */
4969static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans,
4970 struct inode *inode,
4971 struct dentry *parent,
4972 struct super_block *sb,
4973 u64 last_committed)
4974{
4975 int ret = 0;
4976 struct dentry *old_parent = NULL;
4977 struct inode *orig_inode = inode;
4978
4979 /*
4980 * for regular files, if its inode is already on disk, we don't
4981 * have to worry about the parents at all. This is because
4982 * we can use the last_unlink_trans field to record renames
4983 * and other fun in this file.
4984 */
4985 if (S_ISREG(inode->i_mode) &&
4986 BTRFS_I(inode)->generation <= last_committed &&
4987 BTRFS_I(inode)->last_unlink_trans <= last_committed)
4988 goto out;
4989
4990 if (!S_ISDIR(inode->i_mode)) {
4991 if (!parent || d_really_is_negative(parent) || sb != d_inode(parent)->i_sb)
4992 goto out;
4993 inode = d_inode(parent);
4994 }
4995
4996 while (1) {
4997 /*
4998 * If we are logging a directory then we start with our inode,
4999 * not our parents inode, so we need to skipp setting the
5000 * logged_trans so that further down in the log code we don't
5001 * think this inode has already been logged.
5002 */
5003 if (inode != orig_inode)
5004 BTRFS_I(inode)->logged_trans = trans->transid;
5005 smp_mb();
5006
5007 if (btrfs_must_commit_transaction(trans, inode)) {
5008 ret = 1;
5009 break;
5010 }
5011
5012 if (!parent || d_really_is_negative(parent) || sb != d_inode(parent)->i_sb)
5013 break;
5014
5015 if (IS_ROOT(parent))
5016 break;
5017
5018 parent = dget_parent(parent);
5019 dput(old_parent);
5020 old_parent = parent;
5021 inode = d_inode(parent);
5022
5023 }
5024 dput(old_parent);
5025out:
5026 return ret;
5027}
5028
5029struct btrfs_dir_list {
5030 u64 ino;
5031 struct list_head list;
5032};
5033
5034/*
5035 * Log the inodes of the new dentries of a directory. See log_dir_items() for
5036 * details about the why it is needed.
5037 * This is a recursive operation - if an existing dentry corresponds to a
5038 * directory, that directory's new entries are logged too (same behaviour as
5039 * ext3/4, xfs, f2fs, reiserfs, nilfs2). Note that when logging the inodes
5040 * the dentries point to we do not lock their i_mutex, otherwise lockdep
5041 * complains about the following circular lock dependency / possible deadlock:
5042 *
5043 * CPU0 CPU1
5044 * ---- ----
5045 * lock(&type->i_mutex_dir_key#3/2);
5046 * lock(sb_internal#2);
5047 * lock(&type->i_mutex_dir_key#3/2);
5048 * lock(&sb->s_type->i_mutex_key#14);
5049 *
5050 * Where sb_internal is the lock (a counter that works as a lock) acquired by
5051 * sb_start_intwrite() in btrfs_start_transaction().
5052 * Not locking i_mutex of the inodes is still safe because:
5053 *
5054 * 1) For regular files we log with a mode of LOG_INODE_EXISTS. It's possible
5055 * that while logging the inode new references (names) are added or removed
5056 * from the inode, leaving the logged inode item with a link count that does
5057 * not match the number of logged inode reference items. This is fine because
5058 * at log replay time we compute the real number of links and correct the
5059 * link count in the inode item (see replay_one_buffer() and
5060 * link_to_fixup_dir());
5061 *
5062 * 2) For directories we log with a mode of LOG_INODE_ALL. It's possible that
5063 * while logging the inode's items new items with keys BTRFS_DIR_ITEM_KEY and
5064 * BTRFS_DIR_INDEX_KEY are added to fs/subvol tree and the logged inode item
5065 * has a size that doesn't match the sum of the lengths of all the logged
5066 * names. This does not result in a problem because if a dir_item key is
5067 * logged but its matching dir_index key is not logged, at log replay time we
5068 * don't use it to replay the respective name (see replay_one_name()). On the
5069 * other hand if only the dir_index key ends up being logged, the respective
5070 * name is added to the fs/subvol tree with both the dir_item and dir_index
5071 * keys created (see replay_one_name()).
5072 * The directory's inode item with a wrong i_size is not a problem as well,
5073 * since we don't use it at log replay time to set the i_size in the inode
5074 * item of the fs/subvol tree (see overwrite_item()).
5075 */
5076static int log_new_dir_dentries(struct btrfs_trans_handle *trans,
5077 struct btrfs_root *root,
5078 struct inode *start_inode,
5079 struct btrfs_log_ctx *ctx)
5080{
5081 struct btrfs_root *log = root->log_root;
5082 struct btrfs_path *path;
5083 LIST_HEAD(dir_list);
5084 struct btrfs_dir_list *dir_elem;
5085 int ret = 0;
5086
5087 path = btrfs_alloc_path();
5088 if (!path)
5089 return -ENOMEM;
5090
5091 dir_elem = kmalloc(sizeof(*dir_elem), GFP_NOFS);
5092 if (!dir_elem) {
5093 btrfs_free_path(path);
5094 return -ENOMEM;
5095 }
5096 dir_elem->ino = btrfs_ino(start_inode);
5097 list_add_tail(&dir_elem->list, &dir_list);
5098
5099 while (!list_empty(&dir_list)) {
5100 struct extent_buffer *leaf;
5101 struct btrfs_key min_key;
5102 int nritems;
5103 int i;
5104
5105 dir_elem = list_first_entry(&dir_list, struct btrfs_dir_list,
5106 list);
5107 if (ret)
5108 goto next_dir_inode;
5109
5110 min_key.objectid = dir_elem->ino;
5111 min_key.type = BTRFS_DIR_ITEM_KEY;
5112 min_key.offset = 0;
5113again:
5114 btrfs_release_path(path);
5115 ret = btrfs_search_forward(log, &min_key, path, trans->transid);
5116 if (ret < 0) {
5117 goto next_dir_inode;
5118 } else if (ret > 0) {
5119 ret = 0;
5120 goto next_dir_inode;
5121 }
5122
5123process_leaf:
5124 leaf = path->nodes[0];
5125 nritems = btrfs_header_nritems(leaf);
5126 for (i = path->slots[0]; i < nritems; i++) {
5127 struct btrfs_dir_item *di;
5128 struct btrfs_key di_key;
5129 struct inode *di_inode;
5130 struct btrfs_dir_list *new_dir_elem;
5131 int log_mode = LOG_INODE_EXISTS;
5132 int type;
5133
5134 btrfs_item_key_to_cpu(leaf, &min_key, i);
5135 if (min_key.objectid != dir_elem->ino ||
5136 min_key.type != BTRFS_DIR_ITEM_KEY)
5137 goto next_dir_inode;
5138
5139 di = btrfs_item_ptr(leaf, i, struct btrfs_dir_item);
5140 type = btrfs_dir_type(leaf, di);
5141 if (btrfs_dir_transid(leaf, di) < trans->transid &&
5142 type != BTRFS_FT_DIR)
5143 continue;
5144 btrfs_dir_item_key_to_cpu(leaf, di, &di_key);
5145 if (di_key.type == BTRFS_ROOT_ITEM_KEY)
5146 continue;
5147
5148 di_inode = btrfs_iget(root->fs_info->sb, &di_key,
5149 root, NULL);
5150 if (IS_ERR(di_inode)) {
5151 ret = PTR_ERR(di_inode);
5152 goto next_dir_inode;
5153 }
5154
5155 if (btrfs_inode_in_log(di_inode, trans->transid)) {
5156 iput(di_inode);
5157 continue;
5158 }
5159
5160 ctx->log_new_dentries = false;
5161 if (type == BTRFS_FT_DIR)
5162 log_mode = LOG_INODE_ALL;
5163 btrfs_release_path(path);
5164 ret = btrfs_log_inode(trans, root, di_inode,
5165 log_mode, 0, LLONG_MAX, ctx);
5166 if (!ret &&
5167 btrfs_must_commit_transaction(trans, di_inode))
5168 ret = 1;
5169 iput(di_inode);
5170 if (ret)
5171 goto next_dir_inode;
5172 if (ctx->log_new_dentries) {
5173 new_dir_elem = kmalloc(sizeof(*new_dir_elem),
5174 GFP_NOFS);
5175 if (!new_dir_elem) {
5176 ret = -ENOMEM;
5177 goto next_dir_inode;
5178 }
5179 new_dir_elem->ino = di_key.objectid;
5180 list_add_tail(&new_dir_elem->list, &dir_list);
5181 }
5182 break;
5183 }
5184 if (i == nritems) {
5185 ret = btrfs_next_leaf(log, path);
5186 if (ret < 0) {
5187 goto next_dir_inode;
5188 } else if (ret > 0) {
5189 ret = 0;
5190 goto next_dir_inode;
5191 }
5192 goto process_leaf;
5193 }
5194 if (min_key.offset < (u64)-1) {
5195 min_key.offset++;
5196 goto again;
5197 }
5198next_dir_inode:
5199 list_del(&dir_elem->list);
5200 kfree(dir_elem);
5201 }
5202
5203 btrfs_free_path(path);
5204 return ret;
5205}
5206
5207static int btrfs_log_all_parents(struct btrfs_trans_handle *trans,
5208 struct inode *inode,
5209 struct btrfs_log_ctx *ctx)
5210{
5211 int ret;
5212 struct btrfs_path *path;
5213 struct btrfs_key key;
5214 struct btrfs_root *root = BTRFS_I(inode)->root;
5215 const u64 ino = btrfs_ino(inode);
5216
5217 path = btrfs_alloc_path();
5218 if (!path)
5219 return -ENOMEM;
5220 path->skip_locking = 1;
5221 path->search_commit_root = 1;
5222
5223 key.objectid = ino;
5224 key.type = BTRFS_INODE_REF_KEY;
5225 key.offset = 0;
5226 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
5227 if (ret < 0)
5228 goto out;
5229
5230 while (true) {
5231 struct extent_buffer *leaf = path->nodes[0];
5232 int slot = path->slots[0];
5233 u32 cur_offset = 0;
5234 u32 item_size;
5235 unsigned long ptr;
5236
5237 if (slot >= btrfs_header_nritems(leaf)) {
5238 ret = btrfs_next_leaf(root, path);
5239 if (ret < 0)
5240 goto out;
5241 else if (ret > 0)
5242 break;
5243 continue;
5244 }
5245
5246 btrfs_item_key_to_cpu(leaf, &key, slot);
5247 /* BTRFS_INODE_EXTREF_KEY is BTRFS_INODE_REF_KEY + 1 */
5248 if (key.objectid != ino || key.type > BTRFS_INODE_EXTREF_KEY)
5249 break;
5250
5251 item_size = btrfs_item_size_nr(leaf, slot);
5252 ptr = btrfs_item_ptr_offset(leaf, slot);
5253 while (cur_offset < item_size) {
5254 struct btrfs_key inode_key;
5255 struct inode *dir_inode;
5256
5257 inode_key.type = BTRFS_INODE_ITEM_KEY;
5258 inode_key.offset = 0;
5259
5260 if (key.type == BTRFS_INODE_EXTREF_KEY) {
5261 struct btrfs_inode_extref *extref;
5262
5263 extref = (struct btrfs_inode_extref *)
5264 (ptr + cur_offset);
5265 inode_key.objectid = btrfs_inode_extref_parent(
5266 leaf, extref);
5267 cur_offset += sizeof(*extref);
5268 cur_offset += btrfs_inode_extref_name_len(leaf,
5269 extref);
5270 } else {
5271 inode_key.objectid = key.offset;
5272 cur_offset = item_size;
5273 }
5274
5275 dir_inode = btrfs_iget(root->fs_info->sb, &inode_key,
5276 root, NULL);
5277 /* If parent inode was deleted, skip it. */
5278 if (IS_ERR(dir_inode))
5279 continue;
5280
5281 ret = btrfs_log_inode(trans, root, dir_inode,
5282 LOG_INODE_ALL, 0, LLONG_MAX, ctx);
5283 if (!ret &&
5284 btrfs_must_commit_transaction(trans, dir_inode))
5285 ret = 1;
5286 iput(dir_inode);
5287 if (ret)
5288 goto out;
5289 }
5290 path->slots[0]++;
5291 }
5292 ret = 0;
5293out:
5294 btrfs_free_path(path);
5295 return ret;
5296}
5297
5298/*
5299 * helper function around btrfs_log_inode to make sure newly created
5300 * parent directories also end up in the log. A minimal inode and backref
5301 * only logging is done of any parent directories that are older than
5302 * the last committed transaction
5303 */
5304static int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
5305 struct btrfs_root *root, struct inode *inode,
5306 struct dentry *parent,
5307 const loff_t start,
5308 const loff_t end,
5309 int exists_only,
5310 struct btrfs_log_ctx *ctx)
5311{
5312 int inode_only = exists_only ? LOG_INODE_EXISTS : LOG_INODE_ALL;
5313 struct super_block *sb;
5314 struct dentry *old_parent = NULL;
5315 int ret = 0;
5316 u64 last_committed = root->fs_info->last_trans_committed;
5317 bool log_dentries = false;
5318 struct inode *orig_inode = inode;
5319
5320 sb = inode->i_sb;
5321
5322 if (btrfs_test_opt(root, NOTREELOG)) {
5323 ret = 1;
5324 goto end_no_trans;
5325 }
5326
5327 /*
5328 * The prev transaction commit doesn't complete, we need do
5329 * full commit by ourselves.
5330 */
5331 if (root->fs_info->last_trans_log_full_commit >
5332 root->fs_info->last_trans_committed) {
5333 ret = 1;
5334 goto end_no_trans;
5335 }
5336
5337 if (root != BTRFS_I(inode)->root ||
5338 btrfs_root_refs(&root->root_item) == 0) {
5339 ret = 1;
5340 goto end_no_trans;
5341 }
5342
5343 ret = check_parent_dirs_for_sync(trans, inode, parent,
5344 sb, last_committed);
5345 if (ret)
5346 goto end_no_trans;
5347
5348 if (btrfs_inode_in_log(inode, trans->transid)) {
5349 ret = BTRFS_NO_LOG_SYNC;
5350 goto end_no_trans;
5351 }
5352
5353 ret = start_log_trans(trans, root, ctx);
5354 if (ret)
5355 goto end_no_trans;
5356
5357 ret = btrfs_log_inode(trans, root, inode, inode_only, start, end, ctx);
5358 if (ret)
5359 goto end_trans;
5360
5361 /*
5362 * for regular files, if its inode is already on disk, we don't
5363 * have to worry about the parents at all. This is because
5364 * we can use the last_unlink_trans field to record renames
5365 * and other fun in this file.
5366 */
5367 if (S_ISREG(inode->i_mode) &&
5368 BTRFS_I(inode)->generation <= last_committed &&
5369 BTRFS_I(inode)->last_unlink_trans <= last_committed) {
5370 ret = 0;
5371 goto end_trans;
5372 }
5373
5374 if (S_ISDIR(inode->i_mode) && ctx && ctx->log_new_dentries)
5375 log_dentries = true;
5376
5377 /*
5378 * On unlink we must make sure all our current and old parent directores
5379 * inodes are fully logged. This is to prevent leaving dangling
5380 * directory index entries in directories that were our parents but are
5381 * not anymore. Not doing this results in old parent directory being
5382 * impossible to delete after log replay (rmdir will always fail with
5383 * error -ENOTEMPTY).
5384 *
5385 * Example 1:
5386 *
5387 * mkdir testdir
5388 * touch testdir/foo
5389 * ln testdir/foo testdir/bar
5390 * sync
5391 * unlink testdir/bar
5392 * xfs_io -c fsync testdir/foo
5393 * <power failure>
5394 * mount fs, triggers log replay
5395 *
5396 * If we don't log the parent directory (testdir), after log replay the
5397 * directory still has an entry pointing to the file inode using the bar
5398 * name, but a matching BTRFS_INODE_[REF|EXTREF]_KEY does not exist and
5399 * the file inode has a link count of 1.
5400 *
5401 * Example 2:
5402 *
5403 * mkdir testdir
5404 * touch foo
5405 * ln foo testdir/foo2
5406 * ln foo testdir/foo3
5407 * sync
5408 * unlink testdir/foo3
5409 * xfs_io -c fsync foo
5410 * <power failure>
5411 * mount fs, triggers log replay
5412 *
5413 * Similar as the first example, after log replay the parent directory
5414 * testdir still has an entry pointing to the inode file with name foo3
5415 * but the file inode does not have a matching BTRFS_INODE_REF_KEY item
5416 * and has a link count of 2.
5417 */
5418 if (BTRFS_I(inode)->last_unlink_trans > last_committed) {
5419 ret = btrfs_log_all_parents(trans, orig_inode, ctx);
5420 if (ret)
5421 goto end_trans;
5422 }
5423
5424 while (1) {
5425 if (!parent || d_really_is_negative(parent) || sb != d_inode(parent)->i_sb)
5426 break;
5427
5428 inode = d_inode(parent);
5429 if (root != BTRFS_I(inode)->root)
5430 break;
5431
5432 if (BTRFS_I(inode)->generation > last_committed) {
5433 ret = btrfs_log_inode(trans, root, inode,
5434 LOG_INODE_EXISTS,
5435 0, LLONG_MAX, ctx);
5436 if (ret)
5437 goto end_trans;
5438 }
5439 if (IS_ROOT(parent))
5440 break;
5441
5442 parent = dget_parent(parent);
5443 dput(old_parent);
5444 old_parent = parent;
5445 }
5446 if (log_dentries)
5447 ret = log_new_dir_dentries(trans, root, orig_inode, ctx);
5448 else
5449 ret = 0;
5450end_trans:
5451 dput(old_parent);
5452 if (ret < 0) {
5453 btrfs_set_log_full_commit(root->fs_info, trans);
5454 ret = 1;
5455 }
5456
5457 if (ret)
5458 btrfs_remove_log_ctx(root, ctx);
5459 btrfs_end_log_trans(root);
5460end_no_trans:
5461 return ret;
5462}
5463
5464/*
5465 * it is not safe to log dentry if the chunk root has added new
5466 * chunks. This returns 0 if the dentry was logged, and 1 otherwise.
5467 * If this returns 1, you must commit the transaction to safely get your
5468 * data on disk.
5469 */
5470int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
5471 struct btrfs_root *root, struct dentry *dentry,
5472 const loff_t start,
5473 const loff_t end,
5474 struct btrfs_log_ctx *ctx)
5475{
5476 struct dentry *parent = dget_parent(dentry);
5477 int ret;
5478
5479 ret = btrfs_log_inode_parent(trans, root, d_inode(dentry), parent,
5480 start, end, 0, ctx);
5481 dput(parent);
5482
5483 return ret;
5484}
5485
5486/*
5487 * should be called during mount to recover any replay any log trees
5488 * from the FS
5489 */
5490int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
5491{
5492 int ret;
5493 struct btrfs_path *path;
5494 struct btrfs_trans_handle *trans;
5495 struct btrfs_key key;
5496 struct btrfs_key found_key;
5497 struct btrfs_key tmp_key;
5498 struct btrfs_root *log;
5499 struct btrfs_fs_info *fs_info = log_root_tree->fs_info;
5500 struct walk_control wc = {
5501 .process_func = process_one_buffer,
5502 .stage = 0,
5503 };
5504
5505 path = btrfs_alloc_path();
5506 if (!path)
5507 return -ENOMEM;
5508
5509 fs_info->log_root_recovering = 1;
5510
5511 trans = btrfs_start_transaction(fs_info->tree_root, 0);
5512 if (IS_ERR(trans)) {
5513 ret = PTR_ERR(trans);
5514 goto error;
5515 }
5516
5517 wc.trans = trans;
5518 wc.pin = 1;
5519
5520 ret = walk_log_tree(trans, log_root_tree, &wc);
5521 if (ret) {
5522 btrfs_std_error(fs_info, ret, "Failed to pin buffers while "
5523 "recovering log root tree.");
5524 goto error;
5525 }
5526
5527again:
5528 key.objectid = BTRFS_TREE_LOG_OBJECTID;
5529 key.offset = (u64)-1;
5530 key.type = BTRFS_ROOT_ITEM_KEY;
5531
5532 while (1) {
5533 ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
5534
5535 if (ret < 0) {
5536 btrfs_std_error(fs_info, ret,
5537 "Couldn't find tree log root.");
5538 goto error;
5539 }
5540 if (ret > 0) {
5541 if (path->slots[0] == 0)
5542 break;
5543 path->slots[0]--;
5544 }
5545 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
5546 path->slots[0]);
5547 btrfs_release_path(path);
5548 if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
5549 break;
5550
5551 log = btrfs_read_fs_root(log_root_tree, &found_key);
5552 if (IS_ERR(log)) {
5553 ret = PTR_ERR(log);
5554 btrfs_std_error(fs_info, ret,
5555 "Couldn't read tree log root.");
5556 goto error;
5557 }
5558
5559 tmp_key.objectid = found_key.offset;
5560 tmp_key.type = BTRFS_ROOT_ITEM_KEY;
5561 tmp_key.offset = (u64)-1;
5562
5563 wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key);
5564 if (IS_ERR(wc.replay_dest)) {
5565 ret = PTR_ERR(wc.replay_dest);
5566 free_extent_buffer(log->node);
5567 free_extent_buffer(log->commit_root);
5568 kfree(log);
5569 btrfs_std_error(fs_info, ret, "Couldn't read target root "
5570 "for tree log recovery.");
5571 goto error;
5572 }
5573
5574 wc.replay_dest->log_root = log;
5575 btrfs_record_root_in_trans(trans, wc.replay_dest);
5576 ret = walk_log_tree(trans, log, &wc);
5577
5578 if (!ret && wc.stage == LOG_WALK_REPLAY_ALL) {
5579 ret = fixup_inode_link_counts(trans, wc.replay_dest,
5580 path);
5581 }
5582
5583 key.offset = found_key.offset - 1;
5584 wc.replay_dest->log_root = NULL;
5585 free_extent_buffer(log->node);
5586 free_extent_buffer(log->commit_root);
5587 kfree(log);
5588
5589 if (ret)
5590 goto error;
5591
5592 if (found_key.offset == 0)
5593 break;
5594 }
5595 btrfs_release_path(path);
5596
5597 /* step one is to pin it all, step two is to replay just inodes */
5598 if (wc.pin) {
5599 wc.pin = 0;
5600 wc.process_func = replay_one_buffer;
5601 wc.stage = LOG_WALK_REPLAY_INODES;
5602 goto again;
5603 }
5604 /* step three is to replay everything */
5605 if (wc.stage < LOG_WALK_REPLAY_ALL) {
5606 wc.stage++;
5607 goto again;
5608 }
5609
5610 btrfs_free_path(path);
5611
5612 /* step 4: commit the transaction, which also unpins the blocks */
5613 ret = btrfs_commit_transaction(trans, fs_info->tree_root);
5614 if (ret)
5615 return ret;
5616
5617 free_extent_buffer(log_root_tree->node);
5618 log_root_tree->log_root = NULL;
5619 fs_info->log_root_recovering = 0;
5620 kfree(log_root_tree);
5621
5622 return 0;
5623error:
5624 if (wc.trans)
5625 btrfs_end_transaction(wc.trans, fs_info->tree_root);
5626 btrfs_free_path(path);
5627 return ret;
5628}
5629
5630/*
5631 * there are some corner cases where we want to force a full
5632 * commit instead of allowing a directory to be logged.
5633 *
5634 * They revolve around files there were unlinked from the directory, and
5635 * this function updates the parent directory so that a full commit is
5636 * properly done if it is fsync'd later after the unlinks are done.
5637 *
5638 * Must be called before the unlink operations (updates to the subvolume tree,
5639 * inodes, etc) are done.
5640 */
5641void btrfs_record_unlink_dir(struct btrfs_trans_handle *trans,
5642 struct inode *dir, struct inode *inode,
5643 int for_rename)
5644{
5645 /*
5646 * when we're logging a file, if it hasn't been renamed
5647 * or unlinked, and its inode is fully committed on disk,
5648 * we don't have to worry about walking up the directory chain
5649 * to log its parents.
5650 *
5651 * So, we use the last_unlink_trans field to put this transid
5652 * into the file. When the file is logged we check it and
5653 * don't log the parents if the file is fully on disk.
5654 */
5655 if (S_ISREG(inode->i_mode)) {
5656 mutex_lock(&BTRFS_I(inode)->log_mutex);
5657 BTRFS_I(inode)->last_unlink_trans = trans->transid;
5658 mutex_unlock(&BTRFS_I(inode)->log_mutex);
5659 }
5660
5661 /*
5662 * if this directory was already logged any new
5663 * names for this file/dir will get recorded
5664 */
5665 smp_mb();
5666 if (BTRFS_I(dir)->logged_trans == trans->transid)
5667 return;
5668
5669 /*
5670 * if the inode we're about to unlink was logged,
5671 * the log will be properly updated for any new names
5672 */
5673 if (BTRFS_I(inode)->logged_trans == trans->transid)
5674 return;
5675
5676 /*
5677 * when renaming files across directories, if the directory
5678 * there we're unlinking from gets fsync'd later on, there's
5679 * no way to find the destination directory later and fsync it
5680 * properly. So, we have to be conservative and force commits
5681 * so the new name gets discovered.
5682 */
5683 if (for_rename)
5684 goto record;
5685
5686 /* we can safely do the unlink without any special recording */
5687 return;
5688
5689record:
5690 mutex_lock(&BTRFS_I(dir)->log_mutex);
5691 BTRFS_I(dir)->last_unlink_trans = trans->transid;
5692 mutex_unlock(&BTRFS_I(dir)->log_mutex);
5693}
5694
5695/*
5696 * Make sure that if someone attempts to fsync the parent directory of a deleted
5697 * snapshot, it ends up triggering a transaction commit. This is to guarantee
5698 * that after replaying the log tree of the parent directory's root we will not
5699 * see the snapshot anymore and at log replay time we will not see any log tree
5700 * corresponding to the deleted snapshot's root, which could lead to replaying
5701 * it after replaying the log tree of the parent directory (which would replay
5702 * the snapshot delete operation).
5703 *
5704 * Must be called before the actual snapshot destroy operation (updates to the
5705 * parent root and tree of tree roots trees, etc) are done.
5706 */
5707void btrfs_record_snapshot_destroy(struct btrfs_trans_handle *trans,
5708 struct inode *dir)
5709{
5710 mutex_lock(&BTRFS_I(dir)->log_mutex);
5711 BTRFS_I(dir)->last_unlink_trans = trans->transid;
5712 mutex_unlock(&BTRFS_I(dir)->log_mutex);
5713}
5714
5715/*
5716 * Call this after adding a new name for a file and it will properly
5717 * update the log to reflect the new name.
5718 *
5719 * It will return zero if all goes well, and it will return 1 if a
5720 * full transaction commit is required.
5721 */
5722int btrfs_log_new_name(struct btrfs_trans_handle *trans,
5723 struct inode *inode, struct inode *old_dir,
5724 struct dentry *parent)
5725{
5726 struct btrfs_root * root = BTRFS_I(inode)->root;
5727
5728 /*
5729 * this will force the logging code to walk the dentry chain
5730 * up for the file
5731 */
5732 if (S_ISREG(inode->i_mode))
5733 BTRFS_I(inode)->last_unlink_trans = trans->transid;
5734
5735 /*
5736 * if this inode hasn't been logged and directory we're renaming it
5737 * from hasn't been logged, we don't need to log it
5738 */
5739 if (BTRFS_I(inode)->logged_trans <=
5740 root->fs_info->last_trans_committed &&
5741 (!old_dir || BTRFS_I(old_dir)->logged_trans <=
5742 root->fs_info->last_trans_committed))
5743 return 0;
5744
5745 return btrfs_log_inode_parent(trans, root, inode, parent, 0,
5746 LLONG_MAX, 1, NULL);
5747}
5748
1/*
2 * Copyright (C) 2008 Oracle. All rights reserved.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
17 */
18
19#include <linux/sched.h>
20#include <linux/slab.h>
21#include "ctree.h"
22#include "transaction.h"
23#include "disk-io.h"
24#include "locking.h"
25#include "print-tree.h"
26#include "compat.h"
27#include "tree-log.h"
28
29/* magic values for the inode_only field in btrfs_log_inode:
30 *
31 * LOG_INODE_ALL means to log everything
32 * LOG_INODE_EXISTS means to log just enough to recreate the inode
33 * during log replay
34 */
35#define LOG_INODE_ALL 0
36#define LOG_INODE_EXISTS 1
37
38/*
39 * directory trouble cases
40 *
41 * 1) on rename or unlink, if the inode being unlinked isn't in the fsync
42 * log, we must force a full commit before doing an fsync of the directory
43 * where the unlink was done.
44 * ---> record transid of last unlink/rename per directory
45 *
46 * mkdir foo/some_dir
47 * normal commit
48 * rename foo/some_dir foo2/some_dir
49 * mkdir foo/some_dir
50 * fsync foo/some_dir/some_file
51 *
52 * The fsync above will unlink the original some_dir without recording
53 * it in its new location (foo2). After a crash, some_dir will be gone
54 * unless the fsync of some_file forces a full commit
55 *
56 * 2) we must log any new names for any file or dir that is in the fsync
57 * log. ---> check inode while renaming/linking.
58 *
59 * 2a) we must log any new names for any file or dir during rename
60 * when the directory they are being removed from was logged.
61 * ---> check inode and old parent dir during rename
62 *
63 * 2a is actually the more important variant. With the extra logging
64 * a crash might unlink the old name without recreating the new one
65 *
66 * 3) after a crash, we must go through any directories with a link count
67 * of zero and redo the rm -rf
68 *
69 * mkdir f1/foo
70 * normal commit
71 * rm -rf f1/foo
72 * fsync(f1)
73 *
74 * The directory f1 was fully removed from the FS, but fsync was never
75 * called on f1, only its parent dir. After a crash the rm -rf must
76 * be replayed. This must be able to recurse down the entire
77 * directory tree. The inode link count fixup code takes care of the
78 * ugly details.
79 */
80
81/*
82 * stages for the tree walking. The first
83 * stage (0) is to only pin down the blocks we find
84 * the second stage (1) is to make sure that all the inodes
85 * we find in the log are created in the subvolume.
86 *
87 * The last stage is to deal with directories and links and extents
88 * and all the other fun semantics
89 */
90#define LOG_WALK_PIN_ONLY 0
91#define LOG_WALK_REPLAY_INODES 1
92#define LOG_WALK_REPLAY_ALL 2
93
94static int btrfs_log_inode(struct btrfs_trans_handle *trans,
95 struct btrfs_root *root, struct inode *inode,
96 int inode_only);
97static int link_to_fixup_dir(struct btrfs_trans_handle *trans,
98 struct btrfs_root *root,
99 struct btrfs_path *path, u64 objectid);
100static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
101 struct btrfs_root *root,
102 struct btrfs_root *log,
103 struct btrfs_path *path,
104 u64 dirid, int del_all);
105
106/*
107 * tree logging is a special write ahead log used to make sure that
108 * fsyncs and O_SYNCs can happen without doing full tree commits.
109 *
110 * Full tree commits are expensive because they require commonly
111 * modified blocks to be recowed, creating many dirty pages in the
112 * extent tree an 4x-6x higher write load than ext3.
113 *
114 * Instead of doing a tree commit on every fsync, we use the
115 * key ranges and transaction ids to find items for a given file or directory
116 * that have changed in this transaction. Those items are copied into
117 * a special tree (one per subvolume root), that tree is written to disk
118 * and then the fsync is considered complete.
119 *
120 * After a crash, items are copied out of the log-tree back into the
121 * subvolume tree. Any file data extents found are recorded in the extent
122 * allocation tree, and the log-tree freed.
123 *
124 * The log tree is read three times, once to pin down all the extents it is
125 * using in ram and once, once to create all the inodes logged in the tree
126 * and once to do all the other items.
127 */
128
129/*
130 * start a sub transaction and setup the log tree
131 * this increments the log tree writer count to make the people
132 * syncing the tree wait for us to finish
133 */
134static int start_log_trans(struct btrfs_trans_handle *trans,
135 struct btrfs_root *root)
136{
137 int ret;
138 int err = 0;
139
140 mutex_lock(&root->log_mutex);
141 if (root->log_root) {
142 if (!root->log_start_pid) {
143 root->log_start_pid = current->pid;
144 root->log_multiple_pids = false;
145 } else if (root->log_start_pid != current->pid) {
146 root->log_multiple_pids = true;
147 }
148
149 root->log_batch++;
150 atomic_inc(&root->log_writers);
151 mutex_unlock(&root->log_mutex);
152 return 0;
153 }
154 root->log_multiple_pids = false;
155 root->log_start_pid = current->pid;
156 mutex_lock(&root->fs_info->tree_log_mutex);
157 if (!root->fs_info->log_root_tree) {
158 ret = btrfs_init_log_root_tree(trans, root->fs_info);
159 if (ret)
160 err = ret;
161 }
162 if (err == 0 && !root->log_root) {
163 ret = btrfs_add_log_tree(trans, root);
164 if (ret)
165 err = ret;
166 }
167 mutex_unlock(&root->fs_info->tree_log_mutex);
168 root->log_batch++;
169 atomic_inc(&root->log_writers);
170 mutex_unlock(&root->log_mutex);
171 return err;
172}
173
174/*
175 * returns 0 if there was a log transaction running and we were able
176 * to join, or returns -ENOENT if there were not transactions
177 * in progress
178 */
179static int join_running_log_trans(struct btrfs_root *root)
180{
181 int ret = -ENOENT;
182
183 smp_mb();
184 if (!root->log_root)
185 return -ENOENT;
186
187 mutex_lock(&root->log_mutex);
188 if (root->log_root) {
189 ret = 0;
190 atomic_inc(&root->log_writers);
191 }
192 mutex_unlock(&root->log_mutex);
193 return ret;
194}
195
196/*
197 * This either makes the current running log transaction wait
198 * until you call btrfs_end_log_trans() or it makes any future
199 * log transactions wait until you call btrfs_end_log_trans()
200 */
201int btrfs_pin_log_trans(struct btrfs_root *root)
202{
203 int ret = -ENOENT;
204
205 mutex_lock(&root->log_mutex);
206 atomic_inc(&root->log_writers);
207 mutex_unlock(&root->log_mutex);
208 return ret;
209}
210
211/*
212 * indicate we're done making changes to the log tree
213 * and wake up anyone waiting to do a sync
214 */
215void btrfs_end_log_trans(struct btrfs_root *root)
216{
217 if (atomic_dec_and_test(&root->log_writers)) {
218 smp_mb();
219 if (waitqueue_active(&root->log_writer_wait))
220 wake_up(&root->log_writer_wait);
221 }
222}
223
224
225/*
226 * the walk control struct is used to pass state down the chain when
227 * processing the log tree. The stage field tells us which part
228 * of the log tree processing we are currently doing. The others
229 * are state fields used for that specific part
230 */
231struct walk_control {
232 /* should we free the extent on disk when done? This is used
233 * at transaction commit time while freeing a log tree
234 */
235 int free;
236
237 /* should we write out the extent buffer? This is used
238 * while flushing the log tree to disk during a sync
239 */
240 int write;
241
242 /* should we wait for the extent buffer io to finish? Also used
243 * while flushing the log tree to disk for a sync
244 */
245 int wait;
246
247 /* pin only walk, we record which extents on disk belong to the
248 * log trees
249 */
250 int pin;
251
252 /* what stage of the replay code we're currently in */
253 int stage;
254
255 /* the root we are currently replaying */
256 struct btrfs_root *replay_dest;
257
258 /* the trans handle for the current replay */
259 struct btrfs_trans_handle *trans;
260
261 /* the function that gets used to process blocks we find in the
262 * tree. Note the extent_buffer might not be up to date when it is
263 * passed in, and it must be checked or read if you need the data
264 * inside it
265 */
266 int (*process_func)(struct btrfs_root *log, struct extent_buffer *eb,
267 struct walk_control *wc, u64 gen);
268};
269
270/*
271 * process_func used to pin down extents, write them or wait on them
272 */
273static int process_one_buffer(struct btrfs_root *log,
274 struct extent_buffer *eb,
275 struct walk_control *wc, u64 gen)
276{
277 if (wc->pin)
278 btrfs_pin_extent_for_log_replay(wc->trans,
279 log->fs_info->extent_root,
280 eb->start, eb->len);
281
282 if (btrfs_buffer_uptodate(eb, gen, 0)) {
283 if (wc->write)
284 btrfs_write_tree_block(eb);
285 if (wc->wait)
286 btrfs_wait_tree_block_writeback(eb);
287 }
288 return 0;
289}
290
291/*
292 * Item overwrite used by replay and tree logging. eb, slot and key all refer
293 * to the src data we are copying out.
294 *
295 * root is the tree we are copying into, and path is a scratch
296 * path for use in this function (it should be released on entry and
297 * will be released on exit).
298 *
299 * If the key is already in the destination tree the existing item is
300 * overwritten. If the existing item isn't big enough, it is extended.
301 * If it is too large, it is truncated.
302 *
303 * If the key isn't in the destination yet, a new item is inserted.
304 */
305static noinline int overwrite_item(struct btrfs_trans_handle *trans,
306 struct btrfs_root *root,
307 struct btrfs_path *path,
308 struct extent_buffer *eb, int slot,
309 struct btrfs_key *key)
310{
311 int ret;
312 u32 item_size;
313 u64 saved_i_size = 0;
314 int save_old_i_size = 0;
315 unsigned long src_ptr;
316 unsigned long dst_ptr;
317 int overwrite_root = 0;
318
319 if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID)
320 overwrite_root = 1;
321
322 item_size = btrfs_item_size_nr(eb, slot);
323 src_ptr = btrfs_item_ptr_offset(eb, slot);
324
325 /* look for the key in the destination tree */
326 ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
327 if (ret == 0) {
328 char *src_copy;
329 char *dst_copy;
330 u32 dst_size = btrfs_item_size_nr(path->nodes[0],
331 path->slots[0]);
332 if (dst_size != item_size)
333 goto insert;
334
335 if (item_size == 0) {
336 btrfs_release_path(path);
337 return 0;
338 }
339 dst_copy = kmalloc(item_size, GFP_NOFS);
340 src_copy = kmalloc(item_size, GFP_NOFS);
341 if (!dst_copy || !src_copy) {
342 btrfs_release_path(path);
343 kfree(dst_copy);
344 kfree(src_copy);
345 return -ENOMEM;
346 }
347
348 read_extent_buffer(eb, src_copy, src_ptr, item_size);
349
350 dst_ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
351 read_extent_buffer(path->nodes[0], dst_copy, dst_ptr,
352 item_size);
353 ret = memcmp(dst_copy, src_copy, item_size);
354
355 kfree(dst_copy);
356 kfree(src_copy);
357 /*
358 * they have the same contents, just return, this saves
359 * us from cowing blocks in the destination tree and doing
360 * extra writes that may not have been done by a previous
361 * sync
362 */
363 if (ret == 0) {
364 btrfs_release_path(path);
365 return 0;
366 }
367
368 }
369insert:
370 btrfs_release_path(path);
371 /* try to insert the key into the destination tree */
372 ret = btrfs_insert_empty_item(trans, root, path,
373 key, item_size);
374
375 /* make sure any existing item is the correct size */
376 if (ret == -EEXIST) {
377 u32 found_size;
378 found_size = btrfs_item_size_nr(path->nodes[0],
379 path->slots[0]);
380 if (found_size > item_size)
381 btrfs_truncate_item(trans, root, path, item_size, 1);
382 else if (found_size < item_size)
383 btrfs_extend_item(trans, root, path,
384 item_size - found_size);
385 } else if (ret) {
386 return ret;
387 }
388 dst_ptr = btrfs_item_ptr_offset(path->nodes[0],
389 path->slots[0]);
390
391 /* don't overwrite an existing inode if the generation number
392 * was logged as zero. This is done when the tree logging code
393 * is just logging an inode to make sure it exists after recovery.
394 *
395 * Also, don't overwrite i_size on directories during replay.
396 * log replay inserts and removes directory items based on the
397 * state of the tree found in the subvolume, and i_size is modified
398 * as it goes
399 */
400 if (key->type == BTRFS_INODE_ITEM_KEY && ret == -EEXIST) {
401 struct btrfs_inode_item *src_item;
402 struct btrfs_inode_item *dst_item;
403
404 src_item = (struct btrfs_inode_item *)src_ptr;
405 dst_item = (struct btrfs_inode_item *)dst_ptr;
406
407 if (btrfs_inode_generation(eb, src_item) == 0)
408 goto no_copy;
409
410 if (overwrite_root &&
411 S_ISDIR(btrfs_inode_mode(eb, src_item)) &&
412 S_ISDIR(btrfs_inode_mode(path->nodes[0], dst_item))) {
413 save_old_i_size = 1;
414 saved_i_size = btrfs_inode_size(path->nodes[0],
415 dst_item);
416 }
417 }
418
419 copy_extent_buffer(path->nodes[0], eb, dst_ptr,
420 src_ptr, item_size);
421
422 if (save_old_i_size) {
423 struct btrfs_inode_item *dst_item;
424 dst_item = (struct btrfs_inode_item *)dst_ptr;
425 btrfs_set_inode_size(path->nodes[0], dst_item, saved_i_size);
426 }
427
428 /* make sure the generation is filled in */
429 if (key->type == BTRFS_INODE_ITEM_KEY) {
430 struct btrfs_inode_item *dst_item;
431 dst_item = (struct btrfs_inode_item *)dst_ptr;
432 if (btrfs_inode_generation(path->nodes[0], dst_item) == 0) {
433 btrfs_set_inode_generation(path->nodes[0], dst_item,
434 trans->transid);
435 }
436 }
437no_copy:
438 btrfs_mark_buffer_dirty(path->nodes[0]);
439 btrfs_release_path(path);
440 return 0;
441}
442
443/*
444 * simple helper to read an inode off the disk from a given root
445 * This can only be called for subvolume roots and not for the log
446 */
447static noinline struct inode *read_one_inode(struct btrfs_root *root,
448 u64 objectid)
449{
450 struct btrfs_key key;
451 struct inode *inode;
452
453 key.objectid = objectid;
454 key.type = BTRFS_INODE_ITEM_KEY;
455 key.offset = 0;
456 inode = btrfs_iget(root->fs_info->sb, &key, root, NULL);
457 if (IS_ERR(inode)) {
458 inode = NULL;
459 } else if (is_bad_inode(inode)) {
460 iput(inode);
461 inode = NULL;
462 }
463 return inode;
464}
465
466/* replays a single extent in 'eb' at 'slot' with 'key' into the
467 * subvolume 'root'. path is released on entry and should be released
468 * on exit.
469 *
470 * extents in the log tree have not been allocated out of the extent
471 * tree yet. So, this completes the allocation, taking a reference
472 * as required if the extent already exists or creating a new extent
473 * if it isn't in the extent allocation tree yet.
474 *
475 * The extent is inserted into the file, dropping any existing extents
476 * from the file that overlap the new one.
477 */
478static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
479 struct btrfs_root *root,
480 struct btrfs_path *path,
481 struct extent_buffer *eb, int slot,
482 struct btrfs_key *key)
483{
484 int found_type;
485 u64 mask = root->sectorsize - 1;
486 u64 extent_end;
487 u64 alloc_hint;
488 u64 start = key->offset;
489 u64 saved_nbytes;
490 struct btrfs_file_extent_item *item;
491 struct inode *inode = NULL;
492 unsigned long size;
493 int ret = 0;
494
495 item = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
496 found_type = btrfs_file_extent_type(eb, item);
497
498 if (found_type == BTRFS_FILE_EXTENT_REG ||
499 found_type == BTRFS_FILE_EXTENT_PREALLOC)
500 extent_end = start + btrfs_file_extent_num_bytes(eb, item);
501 else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
502 size = btrfs_file_extent_inline_len(eb, item);
503 extent_end = (start + size + mask) & ~mask;
504 } else {
505 ret = 0;
506 goto out;
507 }
508
509 inode = read_one_inode(root, key->objectid);
510 if (!inode) {
511 ret = -EIO;
512 goto out;
513 }
514
515 /*
516 * first check to see if we already have this extent in the
517 * file. This must be done before the btrfs_drop_extents run
518 * so we don't try to drop this extent.
519 */
520 ret = btrfs_lookup_file_extent(trans, root, path, btrfs_ino(inode),
521 start, 0);
522
523 if (ret == 0 &&
524 (found_type == BTRFS_FILE_EXTENT_REG ||
525 found_type == BTRFS_FILE_EXTENT_PREALLOC)) {
526 struct btrfs_file_extent_item cmp1;
527 struct btrfs_file_extent_item cmp2;
528 struct btrfs_file_extent_item *existing;
529 struct extent_buffer *leaf;
530
531 leaf = path->nodes[0];
532 existing = btrfs_item_ptr(leaf, path->slots[0],
533 struct btrfs_file_extent_item);
534
535 read_extent_buffer(eb, &cmp1, (unsigned long)item,
536 sizeof(cmp1));
537 read_extent_buffer(leaf, &cmp2, (unsigned long)existing,
538 sizeof(cmp2));
539
540 /*
541 * we already have a pointer to this exact extent,
542 * we don't have to do anything
543 */
544 if (memcmp(&cmp1, &cmp2, sizeof(cmp1)) == 0) {
545 btrfs_release_path(path);
546 goto out;
547 }
548 }
549 btrfs_release_path(path);
550
551 saved_nbytes = inode_get_bytes(inode);
552 /* drop any overlapping extents */
553 ret = btrfs_drop_extents(trans, inode, start, extent_end,
554 &alloc_hint, 1);
555 BUG_ON(ret);
556
557 if (found_type == BTRFS_FILE_EXTENT_REG ||
558 found_type == BTRFS_FILE_EXTENT_PREALLOC) {
559 u64 offset;
560 unsigned long dest_offset;
561 struct btrfs_key ins;
562
563 ret = btrfs_insert_empty_item(trans, root, path, key,
564 sizeof(*item));
565 BUG_ON(ret);
566 dest_offset = btrfs_item_ptr_offset(path->nodes[0],
567 path->slots[0]);
568 copy_extent_buffer(path->nodes[0], eb, dest_offset,
569 (unsigned long)item, sizeof(*item));
570
571 ins.objectid = btrfs_file_extent_disk_bytenr(eb, item);
572 ins.offset = btrfs_file_extent_disk_num_bytes(eb, item);
573 ins.type = BTRFS_EXTENT_ITEM_KEY;
574 offset = key->offset - btrfs_file_extent_offset(eb, item);
575
576 if (ins.objectid > 0) {
577 u64 csum_start;
578 u64 csum_end;
579 LIST_HEAD(ordered_sums);
580 /*
581 * is this extent already allocated in the extent
582 * allocation tree? If so, just add a reference
583 */
584 ret = btrfs_lookup_extent(root, ins.objectid,
585 ins.offset);
586 if (ret == 0) {
587 ret = btrfs_inc_extent_ref(trans, root,
588 ins.objectid, ins.offset,
589 0, root->root_key.objectid,
590 key->objectid, offset, 0);
591 BUG_ON(ret);
592 } else {
593 /*
594 * insert the extent pointer in the extent
595 * allocation tree
596 */
597 ret = btrfs_alloc_logged_file_extent(trans,
598 root, root->root_key.objectid,
599 key->objectid, offset, &ins);
600 BUG_ON(ret);
601 }
602 btrfs_release_path(path);
603
604 if (btrfs_file_extent_compression(eb, item)) {
605 csum_start = ins.objectid;
606 csum_end = csum_start + ins.offset;
607 } else {
608 csum_start = ins.objectid +
609 btrfs_file_extent_offset(eb, item);
610 csum_end = csum_start +
611 btrfs_file_extent_num_bytes(eb, item);
612 }
613
614 ret = btrfs_lookup_csums_range(root->log_root,
615 csum_start, csum_end - 1,
616 &ordered_sums, 0);
617 BUG_ON(ret);
618 while (!list_empty(&ordered_sums)) {
619 struct btrfs_ordered_sum *sums;
620 sums = list_entry(ordered_sums.next,
621 struct btrfs_ordered_sum,
622 list);
623 ret = btrfs_csum_file_blocks(trans,
624 root->fs_info->csum_root,
625 sums);
626 BUG_ON(ret);
627 list_del(&sums->list);
628 kfree(sums);
629 }
630 } else {
631 btrfs_release_path(path);
632 }
633 } else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
634 /* inline extents are easy, we just overwrite them */
635 ret = overwrite_item(trans, root, path, eb, slot, key);
636 BUG_ON(ret);
637 }
638
639 inode_set_bytes(inode, saved_nbytes);
640 btrfs_update_inode(trans, root, inode);
641out:
642 if (inode)
643 iput(inode);
644 return ret;
645}
646
647/*
648 * when cleaning up conflicts between the directory names in the
649 * subvolume, directory names in the log and directory names in the
650 * inode back references, we may have to unlink inodes from directories.
651 *
652 * This is a helper function to do the unlink of a specific directory
653 * item
654 */
655static noinline int drop_one_dir_item(struct btrfs_trans_handle *trans,
656 struct btrfs_root *root,
657 struct btrfs_path *path,
658 struct inode *dir,
659 struct btrfs_dir_item *di)
660{
661 struct inode *inode;
662 char *name;
663 int name_len;
664 struct extent_buffer *leaf;
665 struct btrfs_key location;
666 int ret;
667
668 leaf = path->nodes[0];
669
670 btrfs_dir_item_key_to_cpu(leaf, di, &location);
671 name_len = btrfs_dir_name_len(leaf, di);
672 name = kmalloc(name_len, GFP_NOFS);
673 if (!name)
674 return -ENOMEM;
675
676 read_extent_buffer(leaf, name, (unsigned long)(di + 1), name_len);
677 btrfs_release_path(path);
678
679 inode = read_one_inode(root, location.objectid);
680 if (!inode) {
681 kfree(name);
682 return -EIO;
683 }
684
685 ret = link_to_fixup_dir(trans, root, path, location.objectid);
686 BUG_ON(ret);
687
688 ret = btrfs_unlink_inode(trans, root, dir, inode, name, name_len);
689 BUG_ON(ret);
690 kfree(name);
691
692 iput(inode);
693
694 btrfs_run_delayed_items(trans, root);
695 return ret;
696}
697
698/*
699 * helper function to see if a given name and sequence number found
700 * in an inode back reference are already in a directory and correctly
701 * point to this inode
702 */
703static noinline int inode_in_dir(struct btrfs_root *root,
704 struct btrfs_path *path,
705 u64 dirid, u64 objectid, u64 index,
706 const char *name, int name_len)
707{
708 struct btrfs_dir_item *di;
709 struct btrfs_key location;
710 int match = 0;
711
712 di = btrfs_lookup_dir_index_item(NULL, root, path, dirid,
713 index, name, name_len, 0);
714 if (di && !IS_ERR(di)) {
715 btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
716 if (location.objectid != objectid)
717 goto out;
718 } else
719 goto out;
720 btrfs_release_path(path);
721
722 di = btrfs_lookup_dir_item(NULL, root, path, dirid, name, name_len, 0);
723 if (di && !IS_ERR(di)) {
724 btrfs_dir_item_key_to_cpu(path->nodes[0], di, &location);
725 if (location.objectid != objectid)
726 goto out;
727 } else
728 goto out;
729 match = 1;
730out:
731 btrfs_release_path(path);
732 return match;
733}
734
735/*
736 * helper function to check a log tree for a named back reference in
737 * an inode. This is used to decide if a back reference that is
738 * found in the subvolume conflicts with what we find in the log.
739 *
740 * inode backreferences may have multiple refs in a single item,
741 * during replay we process one reference at a time, and we don't
742 * want to delete valid links to a file from the subvolume if that
743 * link is also in the log.
744 */
745static noinline int backref_in_log(struct btrfs_root *log,
746 struct btrfs_key *key,
747 char *name, int namelen)
748{
749 struct btrfs_path *path;
750 struct btrfs_inode_ref *ref;
751 unsigned long ptr;
752 unsigned long ptr_end;
753 unsigned long name_ptr;
754 int found_name_len;
755 int item_size;
756 int ret;
757 int match = 0;
758
759 path = btrfs_alloc_path();
760 if (!path)
761 return -ENOMEM;
762
763 ret = btrfs_search_slot(NULL, log, key, path, 0, 0);
764 if (ret != 0)
765 goto out;
766
767 item_size = btrfs_item_size_nr(path->nodes[0], path->slots[0]);
768 ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
769 ptr_end = ptr + item_size;
770 while (ptr < ptr_end) {
771 ref = (struct btrfs_inode_ref *)ptr;
772 found_name_len = btrfs_inode_ref_name_len(path->nodes[0], ref);
773 if (found_name_len == namelen) {
774 name_ptr = (unsigned long)(ref + 1);
775 ret = memcmp_extent_buffer(path->nodes[0], name,
776 name_ptr, namelen);
777 if (ret == 0) {
778 match = 1;
779 goto out;
780 }
781 }
782 ptr = (unsigned long)(ref + 1) + found_name_len;
783 }
784out:
785 btrfs_free_path(path);
786 return match;
787}
788
789
790/*
791 * replay one inode back reference item found in the log tree.
792 * eb, slot and key refer to the buffer and key found in the log tree.
793 * root is the destination we are replaying into, and path is for temp
794 * use by this function. (it should be released on return).
795 */
796static noinline int add_inode_ref(struct btrfs_trans_handle *trans,
797 struct btrfs_root *root,
798 struct btrfs_root *log,
799 struct btrfs_path *path,
800 struct extent_buffer *eb, int slot,
801 struct btrfs_key *key)
802{
803 struct btrfs_inode_ref *ref;
804 struct btrfs_dir_item *di;
805 struct inode *dir;
806 struct inode *inode;
807 unsigned long ref_ptr;
808 unsigned long ref_end;
809 char *name;
810 int namelen;
811 int ret;
812 int search_done = 0;
813
814 /*
815 * it is possible that we didn't log all the parent directories
816 * for a given inode. If we don't find the dir, just don't
817 * copy the back ref in. The link count fixup code will take
818 * care of the rest
819 */
820 dir = read_one_inode(root, key->offset);
821 if (!dir)
822 return -ENOENT;
823
824 inode = read_one_inode(root, key->objectid);
825 if (!inode) {
826 iput(dir);
827 return -EIO;
828 }
829
830 ref_ptr = btrfs_item_ptr_offset(eb, slot);
831 ref_end = ref_ptr + btrfs_item_size_nr(eb, slot);
832
833again:
834 ref = (struct btrfs_inode_ref *)ref_ptr;
835
836 namelen = btrfs_inode_ref_name_len(eb, ref);
837 name = kmalloc(namelen, GFP_NOFS);
838 BUG_ON(!name);
839
840 read_extent_buffer(eb, name, (unsigned long)(ref + 1), namelen);
841
842 /* if we already have a perfect match, we're done */
843 if (inode_in_dir(root, path, btrfs_ino(dir), btrfs_ino(inode),
844 btrfs_inode_ref_index(eb, ref),
845 name, namelen)) {
846 goto out;
847 }
848
849 /*
850 * look for a conflicting back reference in the metadata.
851 * if we find one we have to unlink that name of the file
852 * before we add our new link. Later on, we overwrite any
853 * existing back reference, and we don't want to create
854 * dangling pointers in the directory.
855 */
856
857 if (search_done)
858 goto insert;
859
860 ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
861 if (ret == 0) {
862 char *victim_name;
863 int victim_name_len;
864 struct btrfs_inode_ref *victim_ref;
865 unsigned long ptr;
866 unsigned long ptr_end;
867 struct extent_buffer *leaf = path->nodes[0];
868
869 /* are we trying to overwrite a back ref for the root directory
870 * if so, just jump out, we're done
871 */
872 if (key->objectid == key->offset)
873 goto out_nowrite;
874
875 /* check all the names in this back reference to see
876 * if they are in the log. if so, we allow them to stay
877 * otherwise they must be unlinked as a conflict
878 */
879 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
880 ptr_end = ptr + btrfs_item_size_nr(leaf, path->slots[0]);
881 while (ptr < ptr_end) {
882 victim_ref = (struct btrfs_inode_ref *)ptr;
883 victim_name_len = btrfs_inode_ref_name_len(leaf,
884 victim_ref);
885 victim_name = kmalloc(victim_name_len, GFP_NOFS);
886 BUG_ON(!victim_name);
887
888 read_extent_buffer(leaf, victim_name,
889 (unsigned long)(victim_ref + 1),
890 victim_name_len);
891
892 if (!backref_in_log(log, key, victim_name,
893 victim_name_len)) {
894 btrfs_inc_nlink(inode);
895 btrfs_release_path(path);
896
897 ret = btrfs_unlink_inode(trans, root, dir,
898 inode, victim_name,
899 victim_name_len);
900 btrfs_run_delayed_items(trans, root);
901 }
902 kfree(victim_name);
903 ptr = (unsigned long)(victim_ref + 1) + victim_name_len;
904 }
905 BUG_ON(ret);
906
907 /*
908 * NOTE: we have searched root tree and checked the
909 * coresponding ref, it does not need to check again.
910 */
911 search_done = 1;
912 }
913 btrfs_release_path(path);
914
915 /* look for a conflicting sequence number */
916 di = btrfs_lookup_dir_index_item(trans, root, path, btrfs_ino(dir),
917 btrfs_inode_ref_index(eb, ref),
918 name, namelen, 0);
919 if (di && !IS_ERR(di)) {
920 ret = drop_one_dir_item(trans, root, path, dir, di);
921 BUG_ON(ret);
922 }
923 btrfs_release_path(path);
924
925 /* look for a conflicing name */
926 di = btrfs_lookup_dir_item(trans, root, path, btrfs_ino(dir),
927 name, namelen, 0);
928 if (di && !IS_ERR(di)) {
929 ret = drop_one_dir_item(trans, root, path, dir, di);
930 BUG_ON(ret);
931 }
932 btrfs_release_path(path);
933
934insert:
935 /* insert our name */
936 ret = btrfs_add_link(trans, dir, inode, name, namelen, 0,
937 btrfs_inode_ref_index(eb, ref));
938 BUG_ON(ret);
939
940 btrfs_update_inode(trans, root, inode);
941
942out:
943 ref_ptr = (unsigned long)(ref + 1) + namelen;
944 kfree(name);
945 if (ref_ptr < ref_end)
946 goto again;
947
948 /* finally write the back reference in the inode */
949 ret = overwrite_item(trans, root, path, eb, slot, key);
950 BUG_ON(ret);
951
952out_nowrite:
953 btrfs_release_path(path);
954 iput(dir);
955 iput(inode);
956 return 0;
957}
958
959static int insert_orphan_item(struct btrfs_trans_handle *trans,
960 struct btrfs_root *root, u64 offset)
961{
962 int ret;
963 ret = btrfs_find_orphan_item(root, offset);
964 if (ret > 0)
965 ret = btrfs_insert_orphan_item(trans, root, offset);
966 return ret;
967}
968
969
970/*
971 * There are a few corners where the link count of the file can't
972 * be properly maintained during replay. So, instead of adding
973 * lots of complexity to the log code, we just scan the backrefs
974 * for any file that has been through replay.
975 *
976 * The scan will update the link count on the inode to reflect the
977 * number of back refs found. If it goes down to zero, the iput
978 * will free the inode.
979 */
980static noinline int fixup_inode_link_count(struct btrfs_trans_handle *trans,
981 struct btrfs_root *root,
982 struct inode *inode)
983{
984 struct btrfs_path *path;
985 int ret;
986 struct btrfs_key key;
987 u64 nlink = 0;
988 unsigned long ptr;
989 unsigned long ptr_end;
990 int name_len;
991 u64 ino = btrfs_ino(inode);
992
993 key.objectid = ino;
994 key.type = BTRFS_INODE_REF_KEY;
995 key.offset = (u64)-1;
996
997 path = btrfs_alloc_path();
998 if (!path)
999 return -ENOMEM;
1000
1001 while (1) {
1002 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1003 if (ret < 0)
1004 break;
1005 if (ret > 0) {
1006 if (path->slots[0] == 0)
1007 break;
1008 path->slots[0]--;
1009 }
1010 btrfs_item_key_to_cpu(path->nodes[0], &key,
1011 path->slots[0]);
1012 if (key.objectid != ino ||
1013 key.type != BTRFS_INODE_REF_KEY)
1014 break;
1015 ptr = btrfs_item_ptr_offset(path->nodes[0], path->slots[0]);
1016 ptr_end = ptr + btrfs_item_size_nr(path->nodes[0],
1017 path->slots[0]);
1018 while (ptr < ptr_end) {
1019 struct btrfs_inode_ref *ref;
1020
1021 ref = (struct btrfs_inode_ref *)ptr;
1022 name_len = btrfs_inode_ref_name_len(path->nodes[0],
1023 ref);
1024 ptr = (unsigned long)(ref + 1) + name_len;
1025 nlink++;
1026 }
1027
1028 if (key.offset == 0)
1029 break;
1030 key.offset--;
1031 btrfs_release_path(path);
1032 }
1033 btrfs_release_path(path);
1034 if (nlink != inode->i_nlink) {
1035 set_nlink(inode, nlink);
1036 btrfs_update_inode(trans, root, inode);
1037 }
1038 BTRFS_I(inode)->index_cnt = (u64)-1;
1039
1040 if (inode->i_nlink == 0) {
1041 if (S_ISDIR(inode->i_mode)) {
1042 ret = replay_dir_deletes(trans, root, NULL, path,
1043 ino, 1);
1044 BUG_ON(ret);
1045 }
1046 ret = insert_orphan_item(trans, root, ino);
1047 BUG_ON(ret);
1048 }
1049 btrfs_free_path(path);
1050
1051 return 0;
1052}
1053
1054static noinline int fixup_inode_link_counts(struct btrfs_trans_handle *trans,
1055 struct btrfs_root *root,
1056 struct btrfs_path *path)
1057{
1058 int ret;
1059 struct btrfs_key key;
1060 struct inode *inode;
1061
1062 key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
1063 key.type = BTRFS_ORPHAN_ITEM_KEY;
1064 key.offset = (u64)-1;
1065 while (1) {
1066 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1067 if (ret < 0)
1068 break;
1069
1070 if (ret == 1) {
1071 if (path->slots[0] == 0)
1072 break;
1073 path->slots[0]--;
1074 }
1075
1076 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1077 if (key.objectid != BTRFS_TREE_LOG_FIXUP_OBJECTID ||
1078 key.type != BTRFS_ORPHAN_ITEM_KEY)
1079 break;
1080
1081 ret = btrfs_del_item(trans, root, path);
1082 if (ret)
1083 goto out;
1084
1085 btrfs_release_path(path);
1086 inode = read_one_inode(root, key.offset);
1087 if (!inode)
1088 return -EIO;
1089
1090 ret = fixup_inode_link_count(trans, root, inode);
1091 BUG_ON(ret);
1092
1093 iput(inode);
1094
1095 /*
1096 * fixup on a directory may create new entries,
1097 * make sure we always look for the highset possible
1098 * offset
1099 */
1100 key.offset = (u64)-1;
1101 }
1102 ret = 0;
1103out:
1104 btrfs_release_path(path);
1105 return ret;
1106}
1107
1108
1109/*
1110 * record a given inode in the fixup dir so we can check its link
1111 * count when replay is done. The link count is incremented here
1112 * so the inode won't go away until we check it
1113 */
1114static noinline int link_to_fixup_dir(struct btrfs_trans_handle *trans,
1115 struct btrfs_root *root,
1116 struct btrfs_path *path,
1117 u64 objectid)
1118{
1119 struct btrfs_key key;
1120 int ret = 0;
1121 struct inode *inode;
1122
1123 inode = read_one_inode(root, objectid);
1124 if (!inode)
1125 return -EIO;
1126
1127 key.objectid = BTRFS_TREE_LOG_FIXUP_OBJECTID;
1128 btrfs_set_key_type(&key, BTRFS_ORPHAN_ITEM_KEY);
1129 key.offset = objectid;
1130
1131 ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
1132
1133 btrfs_release_path(path);
1134 if (ret == 0) {
1135 btrfs_inc_nlink(inode);
1136 btrfs_update_inode(trans, root, inode);
1137 } else if (ret == -EEXIST) {
1138 ret = 0;
1139 } else {
1140 BUG();
1141 }
1142 iput(inode);
1143
1144 return ret;
1145}
1146
1147/*
1148 * when replaying the log for a directory, we only insert names
1149 * for inodes that actually exist. This means an fsync on a directory
1150 * does not implicitly fsync all the new files in it
1151 */
1152static noinline int insert_one_name(struct btrfs_trans_handle *trans,
1153 struct btrfs_root *root,
1154 struct btrfs_path *path,
1155 u64 dirid, u64 index,
1156 char *name, int name_len, u8 type,
1157 struct btrfs_key *location)
1158{
1159 struct inode *inode;
1160 struct inode *dir;
1161 int ret;
1162
1163 inode = read_one_inode(root, location->objectid);
1164 if (!inode)
1165 return -ENOENT;
1166
1167 dir = read_one_inode(root, dirid);
1168 if (!dir) {
1169 iput(inode);
1170 return -EIO;
1171 }
1172 ret = btrfs_add_link(trans, dir, inode, name, name_len, 1, index);
1173
1174 /* FIXME, put inode into FIXUP list */
1175
1176 iput(inode);
1177 iput(dir);
1178 return ret;
1179}
1180
1181/*
1182 * take a single entry in a log directory item and replay it into
1183 * the subvolume.
1184 *
1185 * if a conflicting item exists in the subdirectory already,
1186 * the inode it points to is unlinked and put into the link count
1187 * fix up tree.
1188 *
1189 * If a name from the log points to a file or directory that does
1190 * not exist in the FS, it is skipped. fsyncs on directories
1191 * do not force down inodes inside that directory, just changes to the
1192 * names or unlinks in a directory.
1193 */
1194static noinline int replay_one_name(struct btrfs_trans_handle *trans,
1195 struct btrfs_root *root,
1196 struct btrfs_path *path,
1197 struct extent_buffer *eb,
1198 struct btrfs_dir_item *di,
1199 struct btrfs_key *key)
1200{
1201 char *name;
1202 int name_len;
1203 struct btrfs_dir_item *dst_di;
1204 struct btrfs_key found_key;
1205 struct btrfs_key log_key;
1206 struct inode *dir;
1207 u8 log_type;
1208 int exists;
1209 int ret;
1210
1211 dir = read_one_inode(root, key->objectid);
1212 if (!dir)
1213 return -EIO;
1214
1215 name_len = btrfs_dir_name_len(eb, di);
1216 name = kmalloc(name_len, GFP_NOFS);
1217 if (!name)
1218 return -ENOMEM;
1219
1220 log_type = btrfs_dir_type(eb, di);
1221 read_extent_buffer(eb, name, (unsigned long)(di + 1),
1222 name_len);
1223
1224 btrfs_dir_item_key_to_cpu(eb, di, &log_key);
1225 exists = btrfs_lookup_inode(trans, root, path, &log_key, 0);
1226 if (exists == 0)
1227 exists = 1;
1228 else
1229 exists = 0;
1230 btrfs_release_path(path);
1231
1232 if (key->type == BTRFS_DIR_ITEM_KEY) {
1233 dst_di = btrfs_lookup_dir_item(trans, root, path, key->objectid,
1234 name, name_len, 1);
1235 } else if (key->type == BTRFS_DIR_INDEX_KEY) {
1236 dst_di = btrfs_lookup_dir_index_item(trans, root, path,
1237 key->objectid,
1238 key->offset, name,
1239 name_len, 1);
1240 } else {
1241 BUG();
1242 }
1243 if (IS_ERR_OR_NULL(dst_di)) {
1244 /* we need a sequence number to insert, so we only
1245 * do inserts for the BTRFS_DIR_INDEX_KEY types
1246 */
1247 if (key->type != BTRFS_DIR_INDEX_KEY)
1248 goto out;
1249 goto insert;
1250 }
1251
1252 btrfs_dir_item_key_to_cpu(path->nodes[0], dst_di, &found_key);
1253 /* the existing item matches the logged item */
1254 if (found_key.objectid == log_key.objectid &&
1255 found_key.type == log_key.type &&
1256 found_key.offset == log_key.offset &&
1257 btrfs_dir_type(path->nodes[0], dst_di) == log_type) {
1258 goto out;
1259 }
1260
1261 /*
1262 * don't drop the conflicting directory entry if the inode
1263 * for the new entry doesn't exist
1264 */
1265 if (!exists)
1266 goto out;
1267
1268 ret = drop_one_dir_item(trans, root, path, dir, dst_di);
1269 BUG_ON(ret);
1270
1271 if (key->type == BTRFS_DIR_INDEX_KEY)
1272 goto insert;
1273out:
1274 btrfs_release_path(path);
1275 kfree(name);
1276 iput(dir);
1277 return 0;
1278
1279insert:
1280 btrfs_release_path(path);
1281 ret = insert_one_name(trans, root, path, key->objectid, key->offset,
1282 name, name_len, log_type, &log_key);
1283
1284 BUG_ON(ret && ret != -ENOENT);
1285 goto out;
1286}
1287
1288/*
1289 * find all the names in a directory item and reconcile them into
1290 * the subvolume. Only BTRFS_DIR_ITEM_KEY types will have more than
1291 * one name in a directory item, but the same code gets used for
1292 * both directory index types
1293 */
1294static noinline int replay_one_dir_item(struct btrfs_trans_handle *trans,
1295 struct btrfs_root *root,
1296 struct btrfs_path *path,
1297 struct extent_buffer *eb, int slot,
1298 struct btrfs_key *key)
1299{
1300 int ret;
1301 u32 item_size = btrfs_item_size_nr(eb, slot);
1302 struct btrfs_dir_item *di;
1303 int name_len;
1304 unsigned long ptr;
1305 unsigned long ptr_end;
1306
1307 ptr = btrfs_item_ptr_offset(eb, slot);
1308 ptr_end = ptr + item_size;
1309 while (ptr < ptr_end) {
1310 di = (struct btrfs_dir_item *)ptr;
1311 if (verify_dir_item(root, eb, di))
1312 return -EIO;
1313 name_len = btrfs_dir_name_len(eb, di);
1314 ret = replay_one_name(trans, root, path, eb, di, key);
1315 BUG_ON(ret);
1316 ptr = (unsigned long)(di + 1);
1317 ptr += name_len;
1318 }
1319 return 0;
1320}
1321
1322/*
1323 * directory replay has two parts. There are the standard directory
1324 * items in the log copied from the subvolume, and range items
1325 * created in the log while the subvolume was logged.
1326 *
1327 * The range items tell us which parts of the key space the log
1328 * is authoritative for. During replay, if a key in the subvolume
1329 * directory is in a logged range item, but not actually in the log
1330 * that means it was deleted from the directory before the fsync
1331 * and should be removed.
1332 */
1333static noinline int find_dir_range(struct btrfs_root *root,
1334 struct btrfs_path *path,
1335 u64 dirid, int key_type,
1336 u64 *start_ret, u64 *end_ret)
1337{
1338 struct btrfs_key key;
1339 u64 found_end;
1340 struct btrfs_dir_log_item *item;
1341 int ret;
1342 int nritems;
1343
1344 if (*start_ret == (u64)-1)
1345 return 1;
1346
1347 key.objectid = dirid;
1348 key.type = key_type;
1349 key.offset = *start_ret;
1350
1351 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1352 if (ret < 0)
1353 goto out;
1354 if (ret > 0) {
1355 if (path->slots[0] == 0)
1356 goto out;
1357 path->slots[0]--;
1358 }
1359 if (ret != 0)
1360 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1361
1362 if (key.type != key_type || key.objectid != dirid) {
1363 ret = 1;
1364 goto next;
1365 }
1366 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
1367 struct btrfs_dir_log_item);
1368 found_end = btrfs_dir_log_end(path->nodes[0], item);
1369
1370 if (*start_ret >= key.offset && *start_ret <= found_end) {
1371 ret = 0;
1372 *start_ret = key.offset;
1373 *end_ret = found_end;
1374 goto out;
1375 }
1376 ret = 1;
1377next:
1378 /* check the next slot in the tree to see if it is a valid item */
1379 nritems = btrfs_header_nritems(path->nodes[0]);
1380 if (path->slots[0] >= nritems) {
1381 ret = btrfs_next_leaf(root, path);
1382 if (ret)
1383 goto out;
1384 } else {
1385 path->slots[0]++;
1386 }
1387
1388 btrfs_item_key_to_cpu(path->nodes[0], &key, path->slots[0]);
1389
1390 if (key.type != key_type || key.objectid != dirid) {
1391 ret = 1;
1392 goto out;
1393 }
1394 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
1395 struct btrfs_dir_log_item);
1396 found_end = btrfs_dir_log_end(path->nodes[0], item);
1397 *start_ret = key.offset;
1398 *end_ret = found_end;
1399 ret = 0;
1400out:
1401 btrfs_release_path(path);
1402 return ret;
1403}
1404
1405/*
1406 * this looks for a given directory item in the log. If the directory
1407 * item is not in the log, the item is removed and the inode it points
1408 * to is unlinked
1409 */
1410static noinline int check_item_in_log(struct btrfs_trans_handle *trans,
1411 struct btrfs_root *root,
1412 struct btrfs_root *log,
1413 struct btrfs_path *path,
1414 struct btrfs_path *log_path,
1415 struct inode *dir,
1416 struct btrfs_key *dir_key)
1417{
1418 int ret;
1419 struct extent_buffer *eb;
1420 int slot;
1421 u32 item_size;
1422 struct btrfs_dir_item *di;
1423 struct btrfs_dir_item *log_di;
1424 int name_len;
1425 unsigned long ptr;
1426 unsigned long ptr_end;
1427 char *name;
1428 struct inode *inode;
1429 struct btrfs_key location;
1430
1431again:
1432 eb = path->nodes[0];
1433 slot = path->slots[0];
1434 item_size = btrfs_item_size_nr(eb, slot);
1435 ptr = btrfs_item_ptr_offset(eb, slot);
1436 ptr_end = ptr + item_size;
1437 while (ptr < ptr_end) {
1438 di = (struct btrfs_dir_item *)ptr;
1439 if (verify_dir_item(root, eb, di)) {
1440 ret = -EIO;
1441 goto out;
1442 }
1443
1444 name_len = btrfs_dir_name_len(eb, di);
1445 name = kmalloc(name_len, GFP_NOFS);
1446 if (!name) {
1447 ret = -ENOMEM;
1448 goto out;
1449 }
1450 read_extent_buffer(eb, name, (unsigned long)(di + 1),
1451 name_len);
1452 log_di = NULL;
1453 if (log && dir_key->type == BTRFS_DIR_ITEM_KEY) {
1454 log_di = btrfs_lookup_dir_item(trans, log, log_path,
1455 dir_key->objectid,
1456 name, name_len, 0);
1457 } else if (log && dir_key->type == BTRFS_DIR_INDEX_KEY) {
1458 log_di = btrfs_lookup_dir_index_item(trans, log,
1459 log_path,
1460 dir_key->objectid,
1461 dir_key->offset,
1462 name, name_len, 0);
1463 }
1464 if (IS_ERR_OR_NULL(log_di)) {
1465 btrfs_dir_item_key_to_cpu(eb, di, &location);
1466 btrfs_release_path(path);
1467 btrfs_release_path(log_path);
1468 inode = read_one_inode(root, location.objectid);
1469 if (!inode) {
1470 kfree(name);
1471 return -EIO;
1472 }
1473
1474 ret = link_to_fixup_dir(trans, root,
1475 path, location.objectid);
1476 BUG_ON(ret);
1477 btrfs_inc_nlink(inode);
1478 ret = btrfs_unlink_inode(trans, root, dir, inode,
1479 name, name_len);
1480 BUG_ON(ret);
1481
1482 btrfs_run_delayed_items(trans, root);
1483
1484 kfree(name);
1485 iput(inode);
1486
1487 /* there might still be more names under this key
1488 * check and repeat if required
1489 */
1490 ret = btrfs_search_slot(NULL, root, dir_key, path,
1491 0, 0);
1492 if (ret == 0)
1493 goto again;
1494 ret = 0;
1495 goto out;
1496 }
1497 btrfs_release_path(log_path);
1498 kfree(name);
1499
1500 ptr = (unsigned long)(di + 1);
1501 ptr += name_len;
1502 }
1503 ret = 0;
1504out:
1505 btrfs_release_path(path);
1506 btrfs_release_path(log_path);
1507 return ret;
1508}
1509
1510/*
1511 * deletion replay happens before we copy any new directory items
1512 * out of the log or out of backreferences from inodes. It
1513 * scans the log to find ranges of keys that log is authoritative for,
1514 * and then scans the directory to find items in those ranges that are
1515 * not present in the log.
1516 *
1517 * Anything we don't find in the log is unlinked and removed from the
1518 * directory.
1519 */
1520static noinline int replay_dir_deletes(struct btrfs_trans_handle *trans,
1521 struct btrfs_root *root,
1522 struct btrfs_root *log,
1523 struct btrfs_path *path,
1524 u64 dirid, int del_all)
1525{
1526 u64 range_start;
1527 u64 range_end;
1528 int key_type = BTRFS_DIR_LOG_ITEM_KEY;
1529 int ret = 0;
1530 struct btrfs_key dir_key;
1531 struct btrfs_key found_key;
1532 struct btrfs_path *log_path;
1533 struct inode *dir;
1534
1535 dir_key.objectid = dirid;
1536 dir_key.type = BTRFS_DIR_ITEM_KEY;
1537 log_path = btrfs_alloc_path();
1538 if (!log_path)
1539 return -ENOMEM;
1540
1541 dir = read_one_inode(root, dirid);
1542 /* it isn't an error if the inode isn't there, that can happen
1543 * because we replay the deletes before we copy in the inode item
1544 * from the log
1545 */
1546 if (!dir) {
1547 btrfs_free_path(log_path);
1548 return 0;
1549 }
1550again:
1551 range_start = 0;
1552 range_end = 0;
1553 while (1) {
1554 if (del_all)
1555 range_end = (u64)-1;
1556 else {
1557 ret = find_dir_range(log, path, dirid, key_type,
1558 &range_start, &range_end);
1559 if (ret != 0)
1560 break;
1561 }
1562
1563 dir_key.offset = range_start;
1564 while (1) {
1565 int nritems;
1566 ret = btrfs_search_slot(NULL, root, &dir_key, path,
1567 0, 0);
1568 if (ret < 0)
1569 goto out;
1570
1571 nritems = btrfs_header_nritems(path->nodes[0]);
1572 if (path->slots[0] >= nritems) {
1573 ret = btrfs_next_leaf(root, path);
1574 if (ret)
1575 break;
1576 }
1577 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
1578 path->slots[0]);
1579 if (found_key.objectid != dirid ||
1580 found_key.type != dir_key.type)
1581 goto next_type;
1582
1583 if (found_key.offset > range_end)
1584 break;
1585
1586 ret = check_item_in_log(trans, root, log, path,
1587 log_path, dir,
1588 &found_key);
1589 BUG_ON(ret);
1590 if (found_key.offset == (u64)-1)
1591 break;
1592 dir_key.offset = found_key.offset + 1;
1593 }
1594 btrfs_release_path(path);
1595 if (range_end == (u64)-1)
1596 break;
1597 range_start = range_end + 1;
1598 }
1599
1600next_type:
1601 ret = 0;
1602 if (key_type == BTRFS_DIR_LOG_ITEM_KEY) {
1603 key_type = BTRFS_DIR_LOG_INDEX_KEY;
1604 dir_key.type = BTRFS_DIR_INDEX_KEY;
1605 btrfs_release_path(path);
1606 goto again;
1607 }
1608out:
1609 btrfs_release_path(path);
1610 btrfs_free_path(log_path);
1611 iput(dir);
1612 return ret;
1613}
1614
1615/*
1616 * the process_func used to replay items from the log tree. This
1617 * gets called in two different stages. The first stage just looks
1618 * for inodes and makes sure they are all copied into the subvolume.
1619 *
1620 * The second stage copies all the other item types from the log into
1621 * the subvolume. The two stage approach is slower, but gets rid of
1622 * lots of complexity around inodes referencing other inodes that exist
1623 * only in the log (references come from either directory items or inode
1624 * back refs).
1625 */
1626static int replay_one_buffer(struct btrfs_root *log, struct extent_buffer *eb,
1627 struct walk_control *wc, u64 gen)
1628{
1629 int nritems;
1630 struct btrfs_path *path;
1631 struct btrfs_root *root = wc->replay_dest;
1632 struct btrfs_key key;
1633 int level;
1634 int i;
1635 int ret;
1636
1637 ret = btrfs_read_buffer(eb, gen);
1638 if (ret)
1639 return ret;
1640
1641 level = btrfs_header_level(eb);
1642
1643 if (level != 0)
1644 return 0;
1645
1646 path = btrfs_alloc_path();
1647 if (!path)
1648 return -ENOMEM;
1649
1650 nritems = btrfs_header_nritems(eb);
1651 for (i = 0; i < nritems; i++) {
1652 btrfs_item_key_to_cpu(eb, &key, i);
1653
1654 /* inode keys are done during the first stage */
1655 if (key.type == BTRFS_INODE_ITEM_KEY &&
1656 wc->stage == LOG_WALK_REPLAY_INODES) {
1657 struct btrfs_inode_item *inode_item;
1658 u32 mode;
1659
1660 inode_item = btrfs_item_ptr(eb, i,
1661 struct btrfs_inode_item);
1662 mode = btrfs_inode_mode(eb, inode_item);
1663 if (S_ISDIR(mode)) {
1664 ret = replay_dir_deletes(wc->trans,
1665 root, log, path, key.objectid, 0);
1666 BUG_ON(ret);
1667 }
1668 ret = overwrite_item(wc->trans, root, path,
1669 eb, i, &key);
1670 BUG_ON(ret);
1671
1672 /* for regular files, make sure corresponding
1673 * orhpan item exist. extents past the new EOF
1674 * will be truncated later by orphan cleanup.
1675 */
1676 if (S_ISREG(mode)) {
1677 ret = insert_orphan_item(wc->trans, root,
1678 key.objectid);
1679 BUG_ON(ret);
1680 }
1681
1682 ret = link_to_fixup_dir(wc->trans, root,
1683 path, key.objectid);
1684 BUG_ON(ret);
1685 }
1686 if (wc->stage < LOG_WALK_REPLAY_ALL)
1687 continue;
1688
1689 /* these keys are simply copied */
1690 if (key.type == BTRFS_XATTR_ITEM_KEY) {
1691 ret = overwrite_item(wc->trans, root, path,
1692 eb, i, &key);
1693 BUG_ON(ret);
1694 } else if (key.type == BTRFS_INODE_REF_KEY) {
1695 ret = add_inode_ref(wc->trans, root, log, path,
1696 eb, i, &key);
1697 BUG_ON(ret && ret != -ENOENT);
1698 } else if (key.type == BTRFS_EXTENT_DATA_KEY) {
1699 ret = replay_one_extent(wc->trans, root, path,
1700 eb, i, &key);
1701 BUG_ON(ret);
1702 } else if (key.type == BTRFS_DIR_ITEM_KEY ||
1703 key.type == BTRFS_DIR_INDEX_KEY) {
1704 ret = replay_one_dir_item(wc->trans, root, path,
1705 eb, i, &key);
1706 BUG_ON(ret);
1707 }
1708 }
1709 btrfs_free_path(path);
1710 return 0;
1711}
1712
1713static noinline int walk_down_log_tree(struct btrfs_trans_handle *trans,
1714 struct btrfs_root *root,
1715 struct btrfs_path *path, int *level,
1716 struct walk_control *wc)
1717{
1718 u64 root_owner;
1719 u64 bytenr;
1720 u64 ptr_gen;
1721 struct extent_buffer *next;
1722 struct extent_buffer *cur;
1723 struct extent_buffer *parent;
1724 u32 blocksize;
1725 int ret = 0;
1726
1727 WARN_ON(*level < 0);
1728 WARN_ON(*level >= BTRFS_MAX_LEVEL);
1729
1730 while (*level > 0) {
1731 WARN_ON(*level < 0);
1732 WARN_ON(*level >= BTRFS_MAX_LEVEL);
1733 cur = path->nodes[*level];
1734
1735 if (btrfs_header_level(cur) != *level)
1736 WARN_ON(1);
1737
1738 if (path->slots[*level] >=
1739 btrfs_header_nritems(cur))
1740 break;
1741
1742 bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
1743 ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
1744 blocksize = btrfs_level_size(root, *level - 1);
1745
1746 parent = path->nodes[*level];
1747 root_owner = btrfs_header_owner(parent);
1748
1749 next = btrfs_find_create_tree_block(root, bytenr, blocksize);
1750 if (!next)
1751 return -ENOMEM;
1752
1753 if (*level == 1) {
1754 ret = wc->process_func(root, next, wc, ptr_gen);
1755 if (ret)
1756 return ret;
1757
1758 path->slots[*level]++;
1759 if (wc->free) {
1760 ret = btrfs_read_buffer(next, ptr_gen);
1761 if (ret) {
1762 free_extent_buffer(next);
1763 return ret;
1764 }
1765
1766 btrfs_tree_lock(next);
1767 btrfs_set_lock_blocking(next);
1768 clean_tree_block(trans, root, next);
1769 btrfs_wait_tree_block_writeback(next);
1770 btrfs_tree_unlock(next);
1771
1772 WARN_ON(root_owner !=
1773 BTRFS_TREE_LOG_OBJECTID);
1774 ret = btrfs_free_and_pin_reserved_extent(root,
1775 bytenr, blocksize);
1776 BUG_ON(ret); /* -ENOMEM or logic errors */
1777 }
1778 free_extent_buffer(next);
1779 continue;
1780 }
1781 ret = btrfs_read_buffer(next, ptr_gen);
1782 if (ret) {
1783 free_extent_buffer(next);
1784 return ret;
1785 }
1786
1787 WARN_ON(*level <= 0);
1788 if (path->nodes[*level-1])
1789 free_extent_buffer(path->nodes[*level-1]);
1790 path->nodes[*level-1] = next;
1791 *level = btrfs_header_level(next);
1792 path->slots[*level] = 0;
1793 cond_resched();
1794 }
1795 WARN_ON(*level < 0);
1796 WARN_ON(*level >= BTRFS_MAX_LEVEL);
1797
1798 path->slots[*level] = btrfs_header_nritems(path->nodes[*level]);
1799
1800 cond_resched();
1801 return 0;
1802}
1803
1804static noinline int walk_up_log_tree(struct btrfs_trans_handle *trans,
1805 struct btrfs_root *root,
1806 struct btrfs_path *path, int *level,
1807 struct walk_control *wc)
1808{
1809 u64 root_owner;
1810 int i;
1811 int slot;
1812 int ret;
1813
1814 for (i = *level; i < BTRFS_MAX_LEVEL - 1 && path->nodes[i]; i++) {
1815 slot = path->slots[i];
1816 if (slot + 1 < btrfs_header_nritems(path->nodes[i])) {
1817 path->slots[i]++;
1818 *level = i;
1819 WARN_ON(*level == 0);
1820 return 0;
1821 } else {
1822 struct extent_buffer *parent;
1823 if (path->nodes[*level] == root->node)
1824 parent = path->nodes[*level];
1825 else
1826 parent = path->nodes[*level + 1];
1827
1828 root_owner = btrfs_header_owner(parent);
1829 ret = wc->process_func(root, path->nodes[*level], wc,
1830 btrfs_header_generation(path->nodes[*level]));
1831 if (ret)
1832 return ret;
1833
1834 if (wc->free) {
1835 struct extent_buffer *next;
1836
1837 next = path->nodes[*level];
1838
1839 btrfs_tree_lock(next);
1840 btrfs_set_lock_blocking(next);
1841 clean_tree_block(trans, root, next);
1842 btrfs_wait_tree_block_writeback(next);
1843 btrfs_tree_unlock(next);
1844
1845 WARN_ON(root_owner != BTRFS_TREE_LOG_OBJECTID);
1846 ret = btrfs_free_and_pin_reserved_extent(root,
1847 path->nodes[*level]->start,
1848 path->nodes[*level]->len);
1849 BUG_ON(ret);
1850 }
1851 free_extent_buffer(path->nodes[*level]);
1852 path->nodes[*level] = NULL;
1853 *level = i + 1;
1854 }
1855 }
1856 return 1;
1857}
1858
1859/*
1860 * drop the reference count on the tree rooted at 'snap'. This traverses
1861 * the tree freeing any blocks that have a ref count of zero after being
1862 * decremented.
1863 */
1864static int walk_log_tree(struct btrfs_trans_handle *trans,
1865 struct btrfs_root *log, struct walk_control *wc)
1866{
1867 int ret = 0;
1868 int wret;
1869 int level;
1870 struct btrfs_path *path;
1871 int i;
1872 int orig_level;
1873
1874 path = btrfs_alloc_path();
1875 if (!path)
1876 return -ENOMEM;
1877
1878 level = btrfs_header_level(log->node);
1879 orig_level = level;
1880 path->nodes[level] = log->node;
1881 extent_buffer_get(log->node);
1882 path->slots[level] = 0;
1883
1884 while (1) {
1885 wret = walk_down_log_tree(trans, log, path, &level, wc);
1886 if (wret > 0)
1887 break;
1888 if (wret < 0) {
1889 ret = wret;
1890 goto out;
1891 }
1892
1893 wret = walk_up_log_tree(trans, log, path, &level, wc);
1894 if (wret > 0)
1895 break;
1896 if (wret < 0) {
1897 ret = wret;
1898 goto out;
1899 }
1900 }
1901
1902 /* was the root node processed? if not, catch it here */
1903 if (path->nodes[orig_level]) {
1904 ret = wc->process_func(log, path->nodes[orig_level], wc,
1905 btrfs_header_generation(path->nodes[orig_level]));
1906 if (ret)
1907 goto out;
1908 if (wc->free) {
1909 struct extent_buffer *next;
1910
1911 next = path->nodes[orig_level];
1912
1913 btrfs_tree_lock(next);
1914 btrfs_set_lock_blocking(next);
1915 clean_tree_block(trans, log, next);
1916 btrfs_wait_tree_block_writeback(next);
1917 btrfs_tree_unlock(next);
1918
1919 WARN_ON(log->root_key.objectid !=
1920 BTRFS_TREE_LOG_OBJECTID);
1921 ret = btrfs_free_and_pin_reserved_extent(log, next->start,
1922 next->len);
1923 BUG_ON(ret); /* -ENOMEM or logic errors */
1924 }
1925 }
1926
1927out:
1928 for (i = 0; i <= orig_level; i++) {
1929 if (path->nodes[i]) {
1930 free_extent_buffer(path->nodes[i]);
1931 path->nodes[i] = NULL;
1932 }
1933 }
1934 btrfs_free_path(path);
1935 return ret;
1936}
1937
1938/*
1939 * helper function to update the item for a given subvolumes log root
1940 * in the tree of log roots
1941 */
1942static int update_log_root(struct btrfs_trans_handle *trans,
1943 struct btrfs_root *log)
1944{
1945 int ret;
1946
1947 if (log->log_transid == 1) {
1948 /* insert root item on the first sync */
1949 ret = btrfs_insert_root(trans, log->fs_info->log_root_tree,
1950 &log->root_key, &log->root_item);
1951 } else {
1952 ret = btrfs_update_root(trans, log->fs_info->log_root_tree,
1953 &log->root_key, &log->root_item);
1954 }
1955 return ret;
1956}
1957
1958static int wait_log_commit(struct btrfs_trans_handle *trans,
1959 struct btrfs_root *root, unsigned long transid)
1960{
1961 DEFINE_WAIT(wait);
1962 int index = transid % 2;
1963
1964 /*
1965 * we only allow two pending log transactions at a time,
1966 * so we know that if ours is more than 2 older than the
1967 * current transaction, we're done
1968 */
1969 do {
1970 prepare_to_wait(&root->log_commit_wait[index],
1971 &wait, TASK_UNINTERRUPTIBLE);
1972 mutex_unlock(&root->log_mutex);
1973
1974 if (root->fs_info->last_trans_log_full_commit !=
1975 trans->transid && root->log_transid < transid + 2 &&
1976 atomic_read(&root->log_commit[index]))
1977 schedule();
1978
1979 finish_wait(&root->log_commit_wait[index], &wait);
1980 mutex_lock(&root->log_mutex);
1981 } while (root->fs_info->last_trans_log_full_commit !=
1982 trans->transid && root->log_transid < transid + 2 &&
1983 atomic_read(&root->log_commit[index]));
1984 return 0;
1985}
1986
1987static void wait_for_writer(struct btrfs_trans_handle *trans,
1988 struct btrfs_root *root)
1989{
1990 DEFINE_WAIT(wait);
1991 while (root->fs_info->last_trans_log_full_commit !=
1992 trans->transid && atomic_read(&root->log_writers)) {
1993 prepare_to_wait(&root->log_writer_wait,
1994 &wait, TASK_UNINTERRUPTIBLE);
1995 mutex_unlock(&root->log_mutex);
1996 if (root->fs_info->last_trans_log_full_commit !=
1997 trans->transid && atomic_read(&root->log_writers))
1998 schedule();
1999 mutex_lock(&root->log_mutex);
2000 finish_wait(&root->log_writer_wait, &wait);
2001 }
2002}
2003
2004/*
2005 * btrfs_sync_log does sends a given tree log down to the disk and
2006 * updates the super blocks to record it. When this call is done,
2007 * you know that any inodes previously logged are safely on disk only
2008 * if it returns 0.
2009 *
2010 * Any other return value means you need to call btrfs_commit_transaction.
2011 * Some of the edge cases for fsyncing directories that have had unlinks
2012 * or renames done in the past mean that sometimes the only safe
2013 * fsync is to commit the whole FS. When btrfs_sync_log returns -EAGAIN,
2014 * that has happened.
2015 */
2016int btrfs_sync_log(struct btrfs_trans_handle *trans,
2017 struct btrfs_root *root)
2018{
2019 int index1;
2020 int index2;
2021 int mark;
2022 int ret;
2023 struct btrfs_root *log = root->log_root;
2024 struct btrfs_root *log_root_tree = root->fs_info->log_root_tree;
2025 unsigned long log_transid = 0;
2026
2027 mutex_lock(&root->log_mutex);
2028 index1 = root->log_transid % 2;
2029 if (atomic_read(&root->log_commit[index1])) {
2030 wait_log_commit(trans, root, root->log_transid);
2031 mutex_unlock(&root->log_mutex);
2032 return 0;
2033 }
2034 atomic_set(&root->log_commit[index1], 1);
2035
2036 /* wait for previous tree log sync to complete */
2037 if (atomic_read(&root->log_commit[(index1 + 1) % 2]))
2038 wait_log_commit(trans, root, root->log_transid - 1);
2039 while (1) {
2040 unsigned long batch = root->log_batch;
2041 /* when we're on an ssd, just kick the log commit out */
2042 if (!btrfs_test_opt(root, SSD) && root->log_multiple_pids) {
2043 mutex_unlock(&root->log_mutex);
2044 schedule_timeout_uninterruptible(1);
2045 mutex_lock(&root->log_mutex);
2046 }
2047 wait_for_writer(trans, root);
2048 if (batch == root->log_batch)
2049 break;
2050 }
2051
2052 /* bail out if we need to do a full commit */
2053 if (root->fs_info->last_trans_log_full_commit == trans->transid) {
2054 ret = -EAGAIN;
2055 mutex_unlock(&root->log_mutex);
2056 goto out;
2057 }
2058
2059 log_transid = root->log_transid;
2060 if (log_transid % 2 == 0)
2061 mark = EXTENT_DIRTY;
2062 else
2063 mark = EXTENT_NEW;
2064
2065 /* we start IO on all the marked extents here, but we don't actually
2066 * wait for them until later.
2067 */
2068 ret = btrfs_write_marked_extents(log, &log->dirty_log_pages, mark);
2069 if (ret) {
2070 btrfs_abort_transaction(trans, root, ret);
2071 mutex_unlock(&root->log_mutex);
2072 goto out;
2073 }
2074
2075 btrfs_set_root_node(&log->root_item, log->node);
2076
2077 root->log_batch = 0;
2078 root->log_transid++;
2079 log->log_transid = root->log_transid;
2080 root->log_start_pid = 0;
2081 smp_mb();
2082 /*
2083 * IO has been started, blocks of the log tree have WRITTEN flag set
2084 * in their headers. new modifications of the log will be written to
2085 * new positions. so it's safe to allow log writers to go in.
2086 */
2087 mutex_unlock(&root->log_mutex);
2088
2089 mutex_lock(&log_root_tree->log_mutex);
2090 log_root_tree->log_batch++;
2091 atomic_inc(&log_root_tree->log_writers);
2092 mutex_unlock(&log_root_tree->log_mutex);
2093
2094 ret = update_log_root(trans, log);
2095
2096 mutex_lock(&log_root_tree->log_mutex);
2097 if (atomic_dec_and_test(&log_root_tree->log_writers)) {
2098 smp_mb();
2099 if (waitqueue_active(&log_root_tree->log_writer_wait))
2100 wake_up(&log_root_tree->log_writer_wait);
2101 }
2102
2103 if (ret) {
2104 if (ret != -ENOSPC) {
2105 btrfs_abort_transaction(trans, root, ret);
2106 mutex_unlock(&log_root_tree->log_mutex);
2107 goto out;
2108 }
2109 root->fs_info->last_trans_log_full_commit = trans->transid;
2110 btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2111 mutex_unlock(&log_root_tree->log_mutex);
2112 ret = -EAGAIN;
2113 goto out;
2114 }
2115
2116 index2 = log_root_tree->log_transid % 2;
2117 if (atomic_read(&log_root_tree->log_commit[index2])) {
2118 btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2119 wait_log_commit(trans, log_root_tree,
2120 log_root_tree->log_transid);
2121 mutex_unlock(&log_root_tree->log_mutex);
2122 ret = 0;
2123 goto out;
2124 }
2125 atomic_set(&log_root_tree->log_commit[index2], 1);
2126
2127 if (atomic_read(&log_root_tree->log_commit[(index2 + 1) % 2])) {
2128 wait_log_commit(trans, log_root_tree,
2129 log_root_tree->log_transid - 1);
2130 }
2131
2132 wait_for_writer(trans, log_root_tree);
2133
2134 /*
2135 * now that we've moved on to the tree of log tree roots,
2136 * check the full commit flag again
2137 */
2138 if (root->fs_info->last_trans_log_full_commit == trans->transid) {
2139 btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2140 mutex_unlock(&log_root_tree->log_mutex);
2141 ret = -EAGAIN;
2142 goto out_wake_log_root;
2143 }
2144
2145 ret = btrfs_write_and_wait_marked_extents(log_root_tree,
2146 &log_root_tree->dirty_log_pages,
2147 EXTENT_DIRTY | EXTENT_NEW);
2148 if (ret) {
2149 btrfs_abort_transaction(trans, root, ret);
2150 mutex_unlock(&log_root_tree->log_mutex);
2151 goto out_wake_log_root;
2152 }
2153 btrfs_wait_marked_extents(log, &log->dirty_log_pages, mark);
2154
2155 btrfs_set_super_log_root(root->fs_info->super_for_commit,
2156 log_root_tree->node->start);
2157 btrfs_set_super_log_root_level(root->fs_info->super_for_commit,
2158 btrfs_header_level(log_root_tree->node));
2159
2160 log_root_tree->log_batch = 0;
2161 log_root_tree->log_transid++;
2162 smp_mb();
2163
2164 mutex_unlock(&log_root_tree->log_mutex);
2165
2166 /*
2167 * nobody else is going to jump in and write the the ctree
2168 * super here because the log_commit atomic below is protecting
2169 * us. We must be called with a transaction handle pinning
2170 * the running transaction open, so a full commit can't hop
2171 * in and cause problems either.
2172 */
2173 btrfs_scrub_pause_super(root);
2174 write_ctree_super(trans, root->fs_info->tree_root, 1);
2175 btrfs_scrub_continue_super(root);
2176 ret = 0;
2177
2178 mutex_lock(&root->log_mutex);
2179 if (root->last_log_commit < log_transid)
2180 root->last_log_commit = log_transid;
2181 mutex_unlock(&root->log_mutex);
2182
2183out_wake_log_root:
2184 atomic_set(&log_root_tree->log_commit[index2], 0);
2185 smp_mb();
2186 if (waitqueue_active(&log_root_tree->log_commit_wait[index2]))
2187 wake_up(&log_root_tree->log_commit_wait[index2]);
2188out:
2189 atomic_set(&root->log_commit[index1], 0);
2190 smp_mb();
2191 if (waitqueue_active(&root->log_commit_wait[index1]))
2192 wake_up(&root->log_commit_wait[index1]);
2193 return ret;
2194}
2195
2196static void free_log_tree(struct btrfs_trans_handle *trans,
2197 struct btrfs_root *log)
2198{
2199 int ret;
2200 u64 start;
2201 u64 end;
2202 struct walk_control wc = {
2203 .free = 1,
2204 .process_func = process_one_buffer
2205 };
2206
2207 ret = walk_log_tree(trans, log, &wc);
2208 BUG_ON(ret);
2209
2210 while (1) {
2211 ret = find_first_extent_bit(&log->dirty_log_pages,
2212 0, &start, &end, EXTENT_DIRTY | EXTENT_NEW);
2213 if (ret)
2214 break;
2215
2216 clear_extent_bits(&log->dirty_log_pages, start, end,
2217 EXTENT_DIRTY | EXTENT_NEW, GFP_NOFS);
2218 }
2219
2220 free_extent_buffer(log->node);
2221 kfree(log);
2222}
2223
2224/*
2225 * free all the extents used by the tree log. This should be called
2226 * at commit time of the full transaction
2227 */
2228int btrfs_free_log(struct btrfs_trans_handle *trans, struct btrfs_root *root)
2229{
2230 if (root->log_root) {
2231 free_log_tree(trans, root->log_root);
2232 root->log_root = NULL;
2233 }
2234 return 0;
2235}
2236
2237int btrfs_free_log_root_tree(struct btrfs_trans_handle *trans,
2238 struct btrfs_fs_info *fs_info)
2239{
2240 if (fs_info->log_root_tree) {
2241 free_log_tree(trans, fs_info->log_root_tree);
2242 fs_info->log_root_tree = NULL;
2243 }
2244 return 0;
2245}
2246
2247/*
2248 * If both a file and directory are logged, and unlinks or renames are
2249 * mixed in, we have a few interesting corners:
2250 *
2251 * create file X in dir Y
2252 * link file X to X.link in dir Y
2253 * fsync file X
2254 * unlink file X but leave X.link
2255 * fsync dir Y
2256 *
2257 * After a crash we would expect only X.link to exist. But file X
2258 * didn't get fsync'd again so the log has back refs for X and X.link.
2259 *
2260 * We solve this by removing directory entries and inode backrefs from the
2261 * log when a file that was logged in the current transaction is
2262 * unlinked. Any later fsync will include the updated log entries, and
2263 * we'll be able to reconstruct the proper directory items from backrefs.
2264 *
2265 * This optimizations allows us to avoid relogging the entire inode
2266 * or the entire directory.
2267 */
2268int btrfs_del_dir_entries_in_log(struct btrfs_trans_handle *trans,
2269 struct btrfs_root *root,
2270 const char *name, int name_len,
2271 struct inode *dir, u64 index)
2272{
2273 struct btrfs_root *log;
2274 struct btrfs_dir_item *di;
2275 struct btrfs_path *path;
2276 int ret;
2277 int err = 0;
2278 int bytes_del = 0;
2279 u64 dir_ino = btrfs_ino(dir);
2280
2281 if (BTRFS_I(dir)->logged_trans < trans->transid)
2282 return 0;
2283
2284 ret = join_running_log_trans(root);
2285 if (ret)
2286 return 0;
2287
2288 mutex_lock(&BTRFS_I(dir)->log_mutex);
2289
2290 log = root->log_root;
2291 path = btrfs_alloc_path();
2292 if (!path) {
2293 err = -ENOMEM;
2294 goto out_unlock;
2295 }
2296
2297 di = btrfs_lookup_dir_item(trans, log, path, dir_ino,
2298 name, name_len, -1);
2299 if (IS_ERR(di)) {
2300 err = PTR_ERR(di);
2301 goto fail;
2302 }
2303 if (di) {
2304 ret = btrfs_delete_one_dir_name(trans, log, path, di);
2305 bytes_del += name_len;
2306 BUG_ON(ret);
2307 }
2308 btrfs_release_path(path);
2309 di = btrfs_lookup_dir_index_item(trans, log, path, dir_ino,
2310 index, name, name_len, -1);
2311 if (IS_ERR(di)) {
2312 err = PTR_ERR(di);
2313 goto fail;
2314 }
2315 if (di) {
2316 ret = btrfs_delete_one_dir_name(trans, log, path, di);
2317 bytes_del += name_len;
2318 BUG_ON(ret);
2319 }
2320
2321 /* update the directory size in the log to reflect the names
2322 * we have removed
2323 */
2324 if (bytes_del) {
2325 struct btrfs_key key;
2326
2327 key.objectid = dir_ino;
2328 key.offset = 0;
2329 key.type = BTRFS_INODE_ITEM_KEY;
2330 btrfs_release_path(path);
2331
2332 ret = btrfs_search_slot(trans, log, &key, path, 0, 1);
2333 if (ret < 0) {
2334 err = ret;
2335 goto fail;
2336 }
2337 if (ret == 0) {
2338 struct btrfs_inode_item *item;
2339 u64 i_size;
2340
2341 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
2342 struct btrfs_inode_item);
2343 i_size = btrfs_inode_size(path->nodes[0], item);
2344 if (i_size > bytes_del)
2345 i_size -= bytes_del;
2346 else
2347 i_size = 0;
2348 btrfs_set_inode_size(path->nodes[0], item, i_size);
2349 btrfs_mark_buffer_dirty(path->nodes[0]);
2350 } else
2351 ret = 0;
2352 btrfs_release_path(path);
2353 }
2354fail:
2355 btrfs_free_path(path);
2356out_unlock:
2357 mutex_unlock(&BTRFS_I(dir)->log_mutex);
2358 if (ret == -ENOSPC) {
2359 root->fs_info->last_trans_log_full_commit = trans->transid;
2360 ret = 0;
2361 } else if (ret < 0)
2362 btrfs_abort_transaction(trans, root, ret);
2363
2364 btrfs_end_log_trans(root);
2365
2366 return err;
2367}
2368
2369/* see comments for btrfs_del_dir_entries_in_log */
2370int btrfs_del_inode_ref_in_log(struct btrfs_trans_handle *trans,
2371 struct btrfs_root *root,
2372 const char *name, int name_len,
2373 struct inode *inode, u64 dirid)
2374{
2375 struct btrfs_root *log;
2376 u64 index;
2377 int ret;
2378
2379 if (BTRFS_I(inode)->logged_trans < trans->transid)
2380 return 0;
2381
2382 ret = join_running_log_trans(root);
2383 if (ret)
2384 return 0;
2385 log = root->log_root;
2386 mutex_lock(&BTRFS_I(inode)->log_mutex);
2387
2388 ret = btrfs_del_inode_ref(trans, log, name, name_len, btrfs_ino(inode),
2389 dirid, &index);
2390 mutex_unlock(&BTRFS_I(inode)->log_mutex);
2391 if (ret == -ENOSPC) {
2392 root->fs_info->last_trans_log_full_commit = trans->transid;
2393 ret = 0;
2394 } else if (ret < 0 && ret != -ENOENT)
2395 btrfs_abort_transaction(trans, root, ret);
2396 btrfs_end_log_trans(root);
2397
2398 return ret;
2399}
2400
2401/*
2402 * creates a range item in the log for 'dirid'. first_offset and
2403 * last_offset tell us which parts of the key space the log should
2404 * be considered authoritative for.
2405 */
2406static noinline int insert_dir_log_key(struct btrfs_trans_handle *trans,
2407 struct btrfs_root *log,
2408 struct btrfs_path *path,
2409 int key_type, u64 dirid,
2410 u64 first_offset, u64 last_offset)
2411{
2412 int ret;
2413 struct btrfs_key key;
2414 struct btrfs_dir_log_item *item;
2415
2416 key.objectid = dirid;
2417 key.offset = first_offset;
2418 if (key_type == BTRFS_DIR_ITEM_KEY)
2419 key.type = BTRFS_DIR_LOG_ITEM_KEY;
2420 else
2421 key.type = BTRFS_DIR_LOG_INDEX_KEY;
2422 ret = btrfs_insert_empty_item(trans, log, path, &key, sizeof(*item));
2423 if (ret)
2424 return ret;
2425
2426 item = btrfs_item_ptr(path->nodes[0], path->slots[0],
2427 struct btrfs_dir_log_item);
2428 btrfs_set_dir_log_end(path->nodes[0], item, last_offset);
2429 btrfs_mark_buffer_dirty(path->nodes[0]);
2430 btrfs_release_path(path);
2431 return 0;
2432}
2433
2434/*
2435 * log all the items included in the current transaction for a given
2436 * directory. This also creates the range items in the log tree required
2437 * to replay anything deleted before the fsync
2438 */
2439static noinline int log_dir_items(struct btrfs_trans_handle *trans,
2440 struct btrfs_root *root, struct inode *inode,
2441 struct btrfs_path *path,
2442 struct btrfs_path *dst_path, int key_type,
2443 u64 min_offset, u64 *last_offset_ret)
2444{
2445 struct btrfs_key min_key;
2446 struct btrfs_key max_key;
2447 struct btrfs_root *log = root->log_root;
2448 struct extent_buffer *src;
2449 int err = 0;
2450 int ret;
2451 int i;
2452 int nritems;
2453 u64 first_offset = min_offset;
2454 u64 last_offset = (u64)-1;
2455 u64 ino = btrfs_ino(inode);
2456
2457 log = root->log_root;
2458 max_key.objectid = ino;
2459 max_key.offset = (u64)-1;
2460 max_key.type = key_type;
2461
2462 min_key.objectid = ino;
2463 min_key.type = key_type;
2464 min_key.offset = min_offset;
2465
2466 path->keep_locks = 1;
2467
2468 ret = btrfs_search_forward(root, &min_key, &max_key,
2469 path, 0, trans->transid);
2470
2471 /*
2472 * we didn't find anything from this transaction, see if there
2473 * is anything at all
2474 */
2475 if (ret != 0 || min_key.objectid != ino || min_key.type != key_type) {
2476 min_key.objectid = ino;
2477 min_key.type = key_type;
2478 min_key.offset = (u64)-1;
2479 btrfs_release_path(path);
2480 ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
2481 if (ret < 0) {
2482 btrfs_release_path(path);
2483 return ret;
2484 }
2485 ret = btrfs_previous_item(root, path, ino, key_type);
2486
2487 /* if ret == 0 there are items for this type,
2488 * create a range to tell us the last key of this type.
2489 * otherwise, there are no items in this directory after
2490 * *min_offset, and we create a range to indicate that.
2491 */
2492 if (ret == 0) {
2493 struct btrfs_key tmp;
2494 btrfs_item_key_to_cpu(path->nodes[0], &tmp,
2495 path->slots[0]);
2496 if (key_type == tmp.type)
2497 first_offset = max(min_offset, tmp.offset) + 1;
2498 }
2499 goto done;
2500 }
2501
2502 /* go backward to find any previous key */
2503 ret = btrfs_previous_item(root, path, ino, key_type);
2504 if (ret == 0) {
2505 struct btrfs_key tmp;
2506 btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
2507 if (key_type == tmp.type) {
2508 first_offset = tmp.offset;
2509 ret = overwrite_item(trans, log, dst_path,
2510 path->nodes[0], path->slots[0],
2511 &tmp);
2512 if (ret) {
2513 err = ret;
2514 goto done;
2515 }
2516 }
2517 }
2518 btrfs_release_path(path);
2519
2520 /* find the first key from this transaction again */
2521 ret = btrfs_search_slot(NULL, root, &min_key, path, 0, 0);
2522 if (ret != 0) {
2523 WARN_ON(1);
2524 goto done;
2525 }
2526
2527 /*
2528 * we have a block from this transaction, log every item in it
2529 * from our directory
2530 */
2531 while (1) {
2532 struct btrfs_key tmp;
2533 src = path->nodes[0];
2534 nritems = btrfs_header_nritems(src);
2535 for (i = path->slots[0]; i < nritems; i++) {
2536 btrfs_item_key_to_cpu(src, &min_key, i);
2537
2538 if (min_key.objectid != ino || min_key.type != key_type)
2539 goto done;
2540 ret = overwrite_item(trans, log, dst_path, src, i,
2541 &min_key);
2542 if (ret) {
2543 err = ret;
2544 goto done;
2545 }
2546 }
2547 path->slots[0] = nritems;
2548
2549 /*
2550 * look ahead to the next item and see if it is also
2551 * from this directory and from this transaction
2552 */
2553 ret = btrfs_next_leaf(root, path);
2554 if (ret == 1) {
2555 last_offset = (u64)-1;
2556 goto done;
2557 }
2558 btrfs_item_key_to_cpu(path->nodes[0], &tmp, path->slots[0]);
2559 if (tmp.objectid != ino || tmp.type != key_type) {
2560 last_offset = (u64)-1;
2561 goto done;
2562 }
2563 if (btrfs_header_generation(path->nodes[0]) != trans->transid) {
2564 ret = overwrite_item(trans, log, dst_path,
2565 path->nodes[0], path->slots[0],
2566 &tmp);
2567 if (ret)
2568 err = ret;
2569 else
2570 last_offset = tmp.offset;
2571 goto done;
2572 }
2573 }
2574done:
2575 btrfs_release_path(path);
2576 btrfs_release_path(dst_path);
2577
2578 if (err == 0) {
2579 *last_offset_ret = last_offset;
2580 /*
2581 * insert the log range keys to indicate where the log
2582 * is valid
2583 */
2584 ret = insert_dir_log_key(trans, log, path, key_type,
2585 ino, first_offset, last_offset);
2586 if (ret)
2587 err = ret;
2588 }
2589 return err;
2590}
2591
2592/*
2593 * logging directories is very similar to logging inodes, We find all the items
2594 * from the current transaction and write them to the log.
2595 *
2596 * The recovery code scans the directory in the subvolume, and if it finds a
2597 * key in the range logged that is not present in the log tree, then it means
2598 * that dir entry was unlinked during the transaction.
2599 *
2600 * In order for that scan to work, we must include one key smaller than
2601 * the smallest logged by this transaction and one key larger than the largest
2602 * key logged by this transaction.
2603 */
2604static noinline int log_directory_changes(struct btrfs_trans_handle *trans,
2605 struct btrfs_root *root, struct inode *inode,
2606 struct btrfs_path *path,
2607 struct btrfs_path *dst_path)
2608{
2609 u64 min_key;
2610 u64 max_key;
2611 int ret;
2612 int key_type = BTRFS_DIR_ITEM_KEY;
2613
2614again:
2615 min_key = 0;
2616 max_key = 0;
2617 while (1) {
2618 ret = log_dir_items(trans, root, inode, path,
2619 dst_path, key_type, min_key,
2620 &max_key);
2621 if (ret)
2622 return ret;
2623 if (max_key == (u64)-1)
2624 break;
2625 min_key = max_key + 1;
2626 }
2627
2628 if (key_type == BTRFS_DIR_ITEM_KEY) {
2629 key_type = BTRFS_DIR_INDEX_KEY;
2630 goto again;
2631 }
2632 return 0;
2633}
2634
2635/*
2636 * a helper function to drop items from the log before we relog an
2637 * inode. max_key_type indicates the highest item type to remove.
2638 * This cannot be run for file data extents because it does not
2639 * free the extents they point to.
2640 */
2641static int drop_objectid_items(struct btrfs_trans_handle *trans,
2642 struct btrfs_root *log,
2643 struct btrfs_path *path,
2644 u64 objectid, int max_key_type)
2645{
2646 int ret;
2647 struct btrfs_key key;
2648 struct btrfs_key found_key;
2649
2650 key.objectid = objectid;
2651 key.type = max_key_type;
2652 key.offset = (u64)-1;
2653
2654 while (1) {
2655 ret = btrfs_search_slot(trans, log, &key, path, -1, 1);
2656 BUG_ON(ret == 0);
2657 if (ret < 0)
2658 break;
2659
2660 if (path->slots[0] == 0)
2661 break;
2662
2663 path->slots[0]--;
2664 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
2665 path->slots[0]);
2666
2667 if (found_key.objectid != objectid)
2668 break;
2669
2670 ret = btrfs_del_item(trans, log, path);
2671 if (ret)
2672 break;
2673 btrfs_release_path(path);
2674 }
2675 btrfs_release_path(path);
2676 if (ret > 0)
2677 ret = 0;
2678 return ret;
2679}
2680
2681static noinline int copy_items(struct btrfs_trans_handle *trans,
2682 struct btrfs_root *log,
2683 struct btrfs_path *dst_path,
2684 struct extent_buffer *src,
2685 int start_slot, int nr, int inode_only)
2686{
2687 unsigned long src_offset;
2688 unsigned long dst_offset;
2689 struct btrfs_file_extent_item *extent;
2690 struct btrfs_inode_item *inode_item;
2691 int ret;
2692 struct btrfs_key *ins_keys;
2693 u32 *ins_sizes;
2694 char *ins_data;
2695 int i;
2696 struct list_head ordered_sums;
2697
2698 INIT_LIST_HEAD(&ordered_sums);
2699
2700 ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
2701 nr * sizeof(u32), GFP_NOFS);
2702 if (!ins_data)
2703 return -ENOMEM;
2704
2705 ins_sizes = (u32 *)ins_data;
2706 ins_keys = (struct btrfs_key *)(ins_data + nr * sizeof(u32));
2707
2708 for (i = 0; i < nr; i++) {
2709 ins_sizes[i] = btrfs_item_size_nr(src, i + start_slot);
2710 btrfs_item_key_to_cpu(src, ins_keys + i, i + start_slot);
2711 }
2712 ret = btrfs_insert_empty_items(trans, log, dst_path,
2713 ins_keys, ins_sizes, nr);
2714 if (ret) {
2715 kfree(ins_data);
2716 return ret;
2717 }
2718
2719 for (i = 0; i < nr; i++, dst_path->slots[0]++) {
2720 dst_offset = btrfs_item_ptr_offset(dst_path->nodes[0],
2721 dst_path->slots[0]);
2722
2723 src_offset = btrfs_item_ptr_offset(src, start_slot + i);
2724
2725 copy_extent_buffer(dst_path->nodes[0], src, dst_offset,
2726 src_offset, ins_sizes[i]);
2727
2728 if (inode_only == LOG_INODE_EXISTS &&
2729 ins_keys[i].type == BTRFS_INODE_ITEM_KEY) {
2730 inode_item = btrfs_item_ptr(dst_path->nodes[0],
2731 dst_path->slots[0],
2732 struct btrfs_inode_item);
2733 btrfs_set_inode_size(dst_path->nodes[0], inode_item, 0);
2734
2735 /* set the generation to zero so the recover code
2736 * can tell the difference between an logging
2737 * just to say 'this inode exists' and a logging
2738 * to say 'update this inode with these values'
2739 */
2740 btrfs_set_inode_generation(dst_path->nodes[0],
2741 inode_item, 0);
2742 }
2743 /* take a reference on file data extents so that truncates
2744 * or deletes of this inode don't have to relog the inode
2745 * again
2746 */
2747 if (btrfs_key_type(ins_keys + i) == BTRFS_EXTENT_DATA_KEY) {
2748 int found_type;
2749 extent = btrfs_item_ptr(src, start_slot + i,
2750 struct btrfs_file_extent_item);
2751
2752 if (btrfs_file_extent_generation(src, extent) < trans->transid)
2753 continue;
2754
2755 found_type = btrfs_file_extent_type(src, extent);
2756 if (found_type == BTRFS_FILE_EXTENT_REG ||
2757 found_type == BTRFS_FILE_EXTENT_PREALLOC) {
2758 u64 ds, dl, cs, cl;
2759 ds = btrfs_file_extent_disk_bytenr(src,
2760 extent);
2761 /* ds == 0 is a hole */
2762 if (ds == 0)
2763 continue;
2764
2765 dl = btrfs_file_extent_disk_num_bytes(src,
2766 extent);
2767 cs = btrfs_file_extent_offset(src, extent);
2768 cl = btrfs_file_extent_num_bytes(src,
2769 extent);
2770 if (btrfs_file_extent_compression(src,
2771 extent)) {
2772 cs = 0;
2773 cl = dl;
2774 }
2775
2776 ret = btrfs_lookup_csums_range(
2777 log->fs_info->csum_root,
2778 ds + cs, ds + cs + cl - 1,
2779 &ordered_sums, 0);
2780 BUG_ON(ret);
2781 }
2782 }
2783 }
2784
2785 btrfs_mark_buffer_dirty(dst_path->nodes[0]);
2786 btrfs_release_path(dst_path);
2787 kfree(ins_data);
2788
2789 /*
2790 * we have to do this after the loop above to avoid changing the
2791 * log tree while trying to change the log tree.
2792 */
2793 ret = 0;
2794 while (!list_empty(&ordered_sums)) {
2795 struct btrfs_ordered_sum *sums = list_entry(ordered_sums.next,
2796 struct btrfs_ordered_sum,
2797 list);
2798 if (!ret)
2799 ret = btrfs_csum_file_blocks(trans, log, sums);
2800 list_del(&sums->list);
2801 kfree(sums);
2802 }
2803 return ret;
2804}
2805
2806/* log a single inode in the tree log.
2807 * At least one parent directory for this inode must exist in the tree
2808 * or be logged already.
2809 *
2810 * Any items from this inode changed by the current transaction are copied
2811 * to the log tree. An extra reference is taken on any extents in this
2812 * file, allowing us to avoid a whole pile of corner cases around logging
2813 * blocks that have been removed from the tree.
2814 *
2815 * See LOG_INODE_ALL and related defines for a description of what inode_only
2816 * does.
2817 *
2818 * This handles both files and directories.
2819 */
2820static int btrfs_log_inode(struct btrfs_trans_handle *trans,
2821 struct btrfs_root *root, struct inode *inode,
2822 int inode_only)
2823{
2824 struct btrfs_path *path;
2825 struct btrfs_path *dst_path;
2826 struct btrfs_key min_key;
2827 struct btrfs_key max_key;
2828 struct btrfs_root *log = root->log_root;
2829 struct extent_buffer *src = NULL;
2830 int err = 0;
2831 int ret;
2832 int nritems;
2833 int ins_start_slot = 0;
2834 int ins_nr;
2835 u64 ino = btrfs_ino(inode);
2836
2837 log = root->log_root;
2838
2839 path = btrfs_alloc_path();
2840 if (!path)
2841 return -ENOMEM;
2842 dst_path = btrfs_alloc_path();
2843 if (!dst_path) {
2844 btrfs_free_path(path);
2845 return -ENOMEM;
2846 }
2847
2848 min_key.objectid = ino;
2849 min_key.type = BTRFS_INODE_ITEM_KEY;
2850 min_key.offset = 0;
2851
2852 max_key.objectid = ino;
2853
2854 /* today the code can only do partial logging of directories */
2855 if (!S_ISDIR(inode->i_mode))
2856 inode_only = LOG_INODE_ALL;
2857
2858 if (inode_only == LOG_INODE_EXISTS || S_ISDIR(inode->i_mode))
2859 max_key.type = BTRFS_XATTR_ITEM_KEY;
2860 else
2861 max_key.type = (u8)-1;
2862 max_key.offset = (u64)-1;
2863
2864 ret = btrfs_commit_inode_delayed_items(trans, inode);
2865 if (ret) {
2866 btrfs_free_path(path);
2867 btrfs_free_path(dst_path);
2868 return ret;
2869 }
2870
2871 mutex_lock(&BTRFS_I(inode)->log_mutex);
2872
2873 /*
2874 * a brute force approach to making sure we get the most uptodate
2875 * copies of everything.
2876 */
2877 if (S_ISDIR(inode->i_mode)) {
2878 int max_key_type = BTRFS_DIR_LOG_INDEX_KEY;
2879
2880 if (inode_only == LOG_INODE_EXISTS)
2881 max_key_type = BTRFS_XATTR_ITEM_KEY;
2882 ret = drop_objectid_items(trans, log, path, ino, max_key_type);
2883 } else {
2884 ret = btrfs_truncate_inode_items(trans, log, inode, 0, 0);
2885 }
2886 if (ret) {
2887 err = ret;
2888 goto out_unlock;
2889 }
2890 path->keep_locks = 1;
2891
2892 while (1) {
2893 ins_nr = 0;
2894 ret = btrfs_search_forward(root, &min_key, &max_key,
2895 path, 0, trans->transid);
2896 if (ret != 0)
2897 break;
2898again:
2899 /* note, ins_nr might be > 0 here, cleanup outside the loop */
2900 if (min_key.objectid != ino)
2901 break;
2902 if (min_key.type > max_key.type)
2903 break;
2904
2905 src = path->nodes[0];
2906 if (ins_nr && ins_start_slot + ins_nr == path->slots[0]) {
2907 ins_nr++;
2908 goto next_slot;
2909 } else if (!ins_nr) {
2910 ins_start_slot = path->slots[0];
2911 ins_nr = 1;
2912 goto next_slot;
2913 }
2914
2915 ret = copy_items(trans, log, dst_path, src, ins_start_slot,
2916 ins_nr, inode_only);
2917 if (ret) {
2918 err = ret;
2919 goto out_unlock;
2920 }
2921 ins_nr = 1;
2922 ins_start_slot = path->slots[0];
2923next_slot:
2924
2925 nritems = btrfs_header_nritems(path->nodes[0]);
2926 path->slots[0]++;
2927 if (path->slots[0] < nritems) {
2928 btrfs_item_key_to_cpu(path->nodes[0], &min_key,
2929 path->slots[0]);
2930 goto again;
2931 }
2932 if (ins_nr) {
2933 ret = copy_items(trans, log, dst_path, src,
2934 ins_start_slot,
2935 ins_nr, inode_only);
2936 if (ret) {
2937 err = ret;
2938 goto out_unlock;
2939 }
2940 ins_nr = 0;
2941 }
2942 btrfs_release_path(path);
2943
2944 if (min_key.offset < (u64)-1)
2945 min_key.offset++;
2946 else if (min_key.type < (u8)-1)
2947 min_key.type++;
2948 else if (min_key.objectid < (u64)-1)
2949 min_key.objectid++;
2950 else
2951 break;
2952 }
2953 if (ins_nr) {
2954 ret = copy_items(trans, log, dst_path, src,
2955 ins_start_slot,
2956 ins_nr, inode_only);
2957 if (ret) {
2958 err = ret;
2959 goto out_unlock;
2960 }
2961 ins_nr = 0;
2962 }
2963 WARN_ON(ins_nr);
2964 if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {
2965 btrfs_release_path(path);
2966 btrfs_release_path(dst_path);
2967 ret = log_directory_changes(trans, root, inode, path, dst_path);
2968 if (ret) {
2969 err = ret;
2970 goto out_unlock;
2971 }
2972 }
2973 BTRFS_I(inode)->logged_trans = trans->transid;
2974out_unlock:
2975 mutex_unlock(&BTRFS_I(inode)->log_mutex);
2976
2977 btrfs_free_path(path);
2978 btrfs_free_path(dst_path);
2979 return err;
2980}
2981
2982/*
2983 * follow the dentry parent pointers up the chain and see if any
2984 * of the directories in it require a full commit before they can
2985 * be logged. Returns zero if nothing special needs to be done or 1 if
2986 * a full commit is required.
2987 */
2988static noinline int check_parent_dirs_for_sync(struct btrfs_trans_handle *trans,
2989 struct inode *inode,
2990 struct dentry *parent,
2991 struct super_block *sb,
2992 u64 last_committed)
2993{
2994 int ret = 0;
2995 struct btrfs_root *root;
2996 struct dentry *old_parent = NULL;
2997
2998 /*
2999 * for regular files, if its inode is already on disk, we don't
3000 * have to worry about the parents at all. This is because
3001 * we can use the last_unlink_trans field to record renames
3002 * and other fun in this file.
3003 */
3004 if (S_ISREG(inode->i_mode) &&
3005 BTRFS_I(inode)->generation <= last_committed &&
3006 BTRFS_I(inode)->last_unlink_trans <= last_committed)
3007 goto out;
3008
3009 if (!S_ISDIR(inode->i_mode)) {
3010 if (!parent || !parent->d_inode || sb != parent->d_inode->i_sb)
3011 goto out;
3012 inode = parent->d_inode;
3013 }
3014
3015 while (1) {
3016 BTRFS_I(inode)->logged_trans = trans->transid;
3017 smp_mb();
3018
3019 if (BTRFS_I(inode)->last_unlink_trans > last_committed) {
3020 root = BTRFS_I(inode)->root;
3021
3022 /*
3023 * make sure any commits to the log are forced
3024 * to be full commits
3025 */
3026 root->fs_info->last_trans_log_full_commit =
3027 trans->transid;
3028 ret = 1;
3029 break;
3030 }
3031
3032 if (!parent || !parent->d_inode || sb != parent->d_inode->i_sb)
3033 break;
3034
3035 if (IS_ROOT(parent))
3036 break;
3037
3038 parent = dget_parent(parent);
3039 dput(old_parent);
3040 old_parent = parent;
3041 inode = parent->d_inode;
3042
3043 }
3044 dput(old_parent);
3045out:
3046 return ret;
3047}
3048
3049/*
3050 * helper function around btrfs_log_inode to make sure newly created
3051 * parent directories also end up in the log. A minimal inode and backref
3052 * only logging is done of any parent directories that are older than
3053 * the last committed transaction
3054 */
3055int btrfs_log_inode_parent(struct btrfs_trans_handle *trans,
3056 struct btrfs_root *root, struct inode *inode,
3057 struct dentry *parent, int exists_only)
3058{
3059 int inode_only = exists_only ? LOG_INODE_EXISTS : LOG_INODE_ALL;
3060 struct super_block *sb;
3061 struct dentry *old_parent = NULL;
3062 int ret = 0;
3063 u64 last_committed = root->fs_info->last_trans_committed;
3064
3065 sb = inode->i_sb;
3066
3067 if (btrfs_test_opt(root, NOTREELOG)) {
3068 ret = 1;
3069 goto end_no_trans;
3070 }
3071
3072 if (root->fs_info->last_trans_log_full_commit >
3073 root->fs_info->last_trans_committed) {
3074 ret = 1;
3075 goto end_no_trans;
3076 }
3077
3078 if (root != BTRFS_I(inode)->root ||
3079 btrfs_root_refs(&root->root_item) == 0) {
3080 ret = 1;
3081 goto end_no_trans;
3082 }
3083
3084 ret = check_parent_dirs_for_sync(trans, inode, parent,
3085 sb, last_committed);
3086 if (ret)
3087 goto end_no_trans;
3088
3089 if (btrfs_inode_in_log(inode, trans->transid)) {
3090 ret = BTRFS_NO_LOG_SYNC;
3091 goto end_no_trans;
3092 }
3093
3094 ret = start_log_trans(trans, root);
3095 if (ret)
3096 goto end_trans;
3097
3098 ret = btrfs_log_inode(trans, root, inode, inode_only);
3099 if (ret)
3100 goto end_trans;
3101
3102 /*
3103 * for regular files, if its inode is already on disk, we don't
3104 * have to worry about the parents at all. This is because
3105 * we can use the last_unlink_trans field to record renames
3106 * and other fun in this file.
3107 */
3108 if (S_ISREG(inode->i_mode) &&
3109 BTRFS_I(inode)->generation <= last_committed &&
3110 BTRFS_I(inode)->last_unlink_trans <= last_committed) {
3111 ret = 0;
3112 goto end_trans;
3113 }
3114
3115 inode_only = LOG_INODE_EXISTS;
3116 while (1) {
3117 if (!parent || !parent->d_inode || sb != parent->d_inode->i_sb)
3118 break;
3119
3120 inode = parent->d_inode;
3121 if (root != BTRFS_I(inode)->root)
3122 break;
3123
3124 if (BTRFS_I(inode)->generation >
3125 root->fs_info->last_trans_committed) {
3126 ret = btrfs_log_inode(trans, root, inode, inode_only);
3127 if (ret)
3128 goto end_trans;
3129 }
3130 if (IS_ROOT(parent))
3131 break;
3132
3133 parent = dget_parent(parent);
3134 dput(old_parent);
3135 old_parent = parent;
3136 }
3137 ret = 0;
3138end_trans:
3139 dput(old_parent);
3140 if (ret < 0) {
3141 BUG_ON(ret != -ENOSPC);
3142 root->fs_info->last_trans_log_full_commit = trans->transid;
3143 ret = 1;
3144 }
3145 btrfs_end_log_trans(root);
3146end_no_trans:
3147 return ret;
3148}
3149
3150/*
3151 * it is not safe to log dentry if the chunk root has added new
3152 * chunks. This returns 0 if the dentry was logged, and 1 otherwise.
3153 * If this returns 1, you must commit the transaction to safely get your
3154 * data on disk.
3155 */
3156int btrfs_log_dentry_safe(struct btrfs_trans_handle *trans,
3157 struct btrfs_root *root, struct dentry *dentry)
3158{
3159 struct dentry *parent = dget_parent(dentry);
3160 int ret;
3161
3162 ret = btrfs_log_inode_parent(trans, root, dentry->d_inode, parent, 0);
3163 dput(parent);
3164
3165 return ret;
3166}
3167
3168/*
3169 * should be called during mount to recover any replay any log trees
3170 * from the FS
3171 */
3172int btrfs_recover_log_trees(struct btrfs_root *log_root_tree)
3173{
3174 int ret;
3175 struct btrfs_path *path;
3176 struct btrfs_trans_handle *trans;
3177 struct btrfs_key key;
3178 struct btrfs_key found_key;
3179 struct btrfs_key tmp_key;
3180 struct btrfs_root *log;
3181 struct btrfs_fs_info *fs_info = log_root_tree->fs_info;
3182 struct walk_control wc = {
3183 .process_func = process_one_buffer,
3184 .stage = 0,
3185 };
3186
3187 path = btrfs_alloc_path();
3188 if (!path)
3189 return -ENOMEM;
3190
3191 fs_info->log_root_recovering = 1;
3192
3193 trans = btrfs_start_transaction(fs_info->tree_root, 0);
3194 if (IS_ERR(trans)) {
3195 ret = PTR_ERR(trans);
3196 goto error;
3197 }
3198
3199 wc.trans = trans;
3200 wc.pin = 1;
3201
3202 ret = walk_log_tree(trans, log_root_tree, &wc);
3203 if (ret) {
3204 btrfs_error(fs_info, ret, "Failed to pin buffers while "
3205 "recovering log root tree.");
3206 goto error;
3207 }
3208
3209again:
3210 key.objectid = BTRFS_TREE_LOG_OBJECTID;
3211 key.offset = (u64)-1;
3212 btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY);
3213
3214 while (1) {
3215 ret = btrfs_search_slot(NULL, log_root_tree, &key, path, 0, 0);
3216
3217 if (ret < 0) {
3218 btrfs_error(fs_info, ret,
3219 "Couldn't find tree log root.");
3220 goto error;
3221 }
3222 if (ret > 0) {
3223 if (path->slots[0] == 0)
3224 break;
3225 path->slots[0]--;
3226 }
3227 btrfs_item_key_to_cpu(path->nodes[0], &found_key,
3228 path->slots[0]);
3229 btrfs_release_path(path);
3230 if (found_key.objectid != BTRFS_TREE_LOG_OBJECTID)
3231 break;
3232
3233 log = btrfs_read_fs_root_no_radix(log_root_tree,
3234 &found_key);
3235 if (IS_ERR(log)) {
3236 ret = PTR_ERR(log);
3237 btrfs_error(fs_info, ret,
3238 "Couldn't read tree log root.");
3239 goto error;
3240 }
3241
3242 tmp_key.objectid = found_key.offset;
3243 tmp_key.type = BTRFS_ROOT_ITEM_KEY;
3244 tmp_key.offset = (u64)-1;
3245
3246 wc.replay_dest = btrfs_read_fs_root_no_name(fs_info, &tmp_key);
3247 if (IS_ERR(wc.replay_dest)) {
3248 ret = PTR_ERR(wc.replay_dest);
3249 btrfs_error(fs_info, ret, "Couldn't read target root "
3250 "for tree log recovery.");
3251 goto error;
3252 }
3253
3254 wc.replay_dest->log_root = log;
3255 btrfs_record_root_in_trans(trans, wc.replay_dest);
3256 ret = walk_log_tree(trans, log, &wc);
3257 BUG_ON(ret);
3258
3259 if (wc.stage == LOG_WALK_REPLAY_ALL) {
3260 ret = fixup_inode_link_counts(trans, wc.replay_dest,
3261 path);
3262 BUG_ON(ret);
3263 }
3264
3265 key.offset = found_key.offset - 1;
3266 wc.replay_dest->log_root = NULL;
3267 free_extent_buffer(log->node);
3268 free_extent_buffer(log->commit_root);
3269 kfree(log);
3270
3271 if (found_key.offset == 0)
3272 break;
3273 }
3274 btrfs_release_path(path);
3275
3276 /* step one is to pin it all, step two is to replay just inodes */
3277 if (wc.pin) {
3278 wc.pin = 0;
3279 wc.process_func = replay_one_buffer;
3280 wc.stage = LOG_WALK_REPLAY_INODES;
3281 goto again;
3282 }
3283 /* step three is to replay everything */
3284 if (wc.stage < LOG_WALK_REPLAY_ALL) {
3285 wc.stage++;
3286 goto again;
3287 }
3288
3289 btrfs_free_path(path);
3290
3291 free_extent_buffer(log_root_tree->node);
3292 log_root_tree->log_root = NULL;
3293 fs_info->log_root_recovering = 0;
3294
3295 /* step 4: commit the transaction, which also unpins the blocks */
3296 btrfs_commit_transaction(trans, fs_info->tree_root);
3297
3298 kfree(log_root_tree);
3299 return 0;
3300
3301error:
3302 btrfs_free_path(path);
3303 return ret;
3304}
3305
3306/*
3307 * there are some corner cases where we want to force a full
3308 * commit instead of allowing a directory to be logged.
3309 *
3310 * They revolve around files there were unlinked from the directory, and
3311 * this function updates the parent directory so that a full commit is
3312 * properly done if it is fsync'd later after the unlinks are done.
3313 */
3314void btrfs_record_unlink_dir(struct btrfs_trans_handle *trans,
3315 struct inode *dir, struct inode *inode,
3316 int for_rename)
3317{
3318 /*
3319 * when we're logging a file, if it hasn't been renamed
3320 * or unlinked, and its inode is fully committed on disk,
3321 * we don't have to worry about walking up the directory chain
3322 * to log its parents.
3323 *
3324 * So, we use the last_unlink_trans field to put this transid
3325 * into the file. When the file is logged we check it and
3326 * don't log the parents if the file is fully on disk.
3327 */
3328 if (S_ISREG(inode->i_mode))
3329 BTRFS_I(inode)->last_unlink_trans = trans->transid;
3330
3331 /*
3332 * if this directory was already logged any new
3333 * names for this file/dir will get recorded
3334 */
3335 smp_mb();
3336 if (BTRFS_I(dir)->logged_trans == trans->transid)
3337 return;
3338
3339 /*
3340 * if the inode we're about to unlink was logged,
3341 * the log will be properly updated for any new names
3342 */
3343 if (BTRFS_I(inode)->logged_trans == trans->transid)
3344 return;
3345
3346 /*
3347 * when renaming files across directories, if the directory
3348 * there we're unlinking from gets fsync'd later on, there's
3349 * no way to find the destination directory later and fsync it
3350 * properly. So, we have to be conservative and force commits
3351 * so the new name gets discovered.
3352 */
3353 if (for_rename)
3354 goto record;
3355
3356 /* we can safely do the unlink without any special recording */
3357 return;
3358
3359record:
3360 BTRFS_I(dir)->last_unlink_trans = trans->transid;
3361}
3362
3363/*
3364 * Call this after adding a new name for a file and it will properly
3365 * update the log to reflect the new name.
3366 *
3367 * It will return zero if all goes well, and it will return 1 if a
3368 * full transaction commit is required.
3369 */
3370int btrfs_log_new_name(struct btrfs_trans_handle *trans,
3371 struct inode *inode, struct inode *old_dir,
3372 struct dentry *parent)
3373{
3374 struct btrfs_root * root = BTRFS_I(inode)->root;
3375
3376 /*
3377 * this will force the logging code to walk the dentry chain
3378 * up for the file
3379 */
3380 if (S_ISREG(inode->i_mode))
3381 BTRFS_I(inode)->last_unlink_trans = trans->transid;
3382
3383 /*
3384 * if this inode hasn't been logged and directory we're renaming it
3385 * from hasn't been logged, we don't need to log it
3386 */
3387 if (BTRFS_I(inode)->logged_trans <=
3388 root->fs_info->last_trans_committed &&
3389 (!old_dir || BTRFS_I(old_dir)->logged_trans <=
3390 root->fs_info->last_trans_committed))
3391 return 0;
3392
3393 return btrfs_log_inode_parent(trans, root, inode, parent, 1);
3394}
3395